WO2011027557A1 - 位置校正情報収集装置、位置校正情報収集方法、及び位置校正情報収集プログラム - Google Patents
位置校正情報収集装置、位置校正情報収集方法、及び位置校正情報収集プログラム Download PDFInfo
- Publication number
- WO2011027557A1 WO2011027557A1 PCT/JP2010/005397 JP2010005397W WO2011027557A1 WO 2011027557 A1 WO2011027557 A1 WO 2011027557A1 JP 2010005397 W JP2010005397 W JP 2010005397W WO 2011027557 A1 WO2011027557 A1 WO 2011027557A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- person
- entrance
- exit
- detected
- information
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/35—Categorising the entire scene, e.g. birthday party or wedding scene
- G06V20/36—Indoor scenes
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S2205/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S2205/01—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations specially adapted for specific applications
- G01S2205/02—Indoor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S2205/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S2205/01—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations specially adapted for specific applications
- G01S2205/09—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations specially adapted for specific applications for tracking people
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30232—Surveillance
Definitions
- the present invention relates to a position calibration information collection apparatus, a position calibration information collection method, and a position calibration information collection program that can perform calibration related to the position of an observation apparatus without using a marker prepared in advance.
- a camera or, recently, an ultra-wideband wireless [UWB (Ultra Wide Band)] tag may be used as an observation device for monitoring a person.
- UWB Ultra Wide Band
- a tag reader marker is a position tag that stores position information.
- An example of a camera marker is a two-dimensional barcode in which position information is stored.
- Patent Document 1 there is a technique for estimating the self-position of a robot using a known object existing in the environment without using an artificial marker.
- Patent Document 1 requires the robot to learn the three-dimensional shape of a known object in advance, and substitutes a known object that is not an artificial object as a marker. Moreover, the technique of patent document 1 is a technique which cannot be applied to a UWB tag that cannot identify a three-dimensional shape.
- an object of the present invention is to provide a position calibration information collecting apparatus, a position calibration information collecting method, and a position calibration information collecting program capable of performing calibration related to the position of an observation apparatus that monitors a person without using a marker. There is to do.
- the present invention is configured as follows.
- an observation device for acquiring each as observation information
- First detection position estimation means for estimating a position where the person is first detected in the local coordinate space, based on the feature information of the person, the local coordinates, and the time observed by the observation device
- An entrance / exit position estimation unit that estimates local coordinates of an entrance / exit position of the entrance / exit in the local coordinate space based on the position of the person detected for the first time in the local coordinate space, estimated by the initial detection position estimation unit
- Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means
- Position calibration information calculating means for calculating A position calibration information collecting device is provided.
- the position calibration information collecting device Based on the position calibration information calculated based on the global coordinates of the entrance / exit position of the entrance / exit and the local coordinates of the entrance / exit position installed in the global space by the position calibration information calculating device of the position calibration information collecting apparatus.
- the feature information of the person existing in the environment having the doorway, the local coordinates of the position where the person is detected in the local coordinate space of the environment, and the time when the person is detected And an observation device for acquiring each as observation information, Based on the feature information of the person, the local coordinates, and the time observed by the observation device, a last detected position estimation unit that estimates a position where the person was last detected in the local coordinate space; Entry / exit position estimation means for estimating local coordinates of the entrance / exit position of the entrance / exit in the local coordinate space based on the position where the person was last detected in the local coordinate space, estimated by the end detection position estimation means
- Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means
- Position calibration information calculating means for calculating A position calibration information collecting device is provided.
- characteristic information of a person existing in an environment having an entrance / exit local coordinates of a position where the person is detected in the local coordinate space of the environment, and a time when the person is detected are obtained as observation information by the observation device, Based on the feature information of the person, the local coordinates, and the time observed by the observation device, a position at which the person is first detected in the local coordinate space is estimated by an initial detection position estimation unit, Based on the position where the person is first detected in the local coordinate space estimated by the initial detection position estimating means, the local coordinates of the entrance / exit position of the entrance / exit in the local coordinate space are estimated by the entrance / exit position estimating means, Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means Is calculated by the position calibration information calculation means, A position calibration information collecting method is
- a computer The observation device acquires characteristic information of a person existing in an environment having an entrance, local coordinates of the position where the person is detected in the local coordinate space of the environment, and time when the person is detected as observation information. Function to Based on the feature information of the person, the local coordinates, and the time observed by the observation device, a function of estimating a position at which the person is first detected in the local coordinate space by an initial detection position estimating unit; A function for estimating the local coordinates of the entrance / exit position of the entrance / exit in the local coordinate space by the entrance / exit position estimation unit based on the position where the person is first detected in the local coordinate space, estimated by the initial detection position estimation unit.
- Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means
- a function for calculating the position calibration information calculation means A position calibration information collection program for realizing the above is provided.
- the position at which a person is detected for the first time or the position at which a person is last detected can be estimated as the position of an entrance to a closed environment, and the absolute position in the global coordinate system is specified. be able to. Therefore, it becomes possible to calculate calibration information related to the position of the observation apparatus by detecting a person without using a marker.
- FIG. 1 is a block diagram showing a configuration of a position calibration information collecting apparatus according to the first embodiment of the present invention.
- FIG. 2 is a diagram illustrating a room as a living space that is an observation target in the position calibration information collection device according to the first embodiment of the present invention
- FIG. 3A is a diagram showing an example of human detection history information by a camera stored in the human detection history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 3B is a diagram showing an example of human detection history information by the Ultra Wide Band tag reader stored in the human detection history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 4A is a diagram illustrating an example of information of a person's first detection history by a camera stored in a first detection position history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 4B is a diagram illustrating an example of information on the first detection history of the person by the Ultra Wide Band tag reader stored in the first detection position history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 5 is a diagram showing an example of an environment map of the position calibration information collecting apparatus according to the first embodiment of the present invention.
- FIG. 6A is a diagram showing an example of information of a person's initial detection history by a camera stored in the initial detection position history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 6B is a diagram showing an example of clustering information on the first detection history of a person by a camera stored in the first detection position history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 6C is a diagram illustrating an estimation example of the entrance / exit position estimation unit of the position calibration information collection device according to the first embodiment of the present invention
- FIG. 7A is a diagram showing an example of human detection history information by a camera stored in the human detection history database of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 7B is a diagram showing an example of fattening a human detection position by a camera stored in the human detection history database of the position calibration information collection device according to the first embodiment of the present invention
- FIG. 8 is a flowchart showing the overall processing of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 9 is a diagram showing an example of a tag reader position detection method in the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 10 is a block diagram showing a configuration of the position calibration information collecting apparatus according to the first embodiment of the present invention
- FIG. 11 is a flowchart showing a calibration parameter acquisition process using a person's stay position and an inaccessible position of the position calibration information collection apparatus according to the first embodiment of the present invention
- FIG. 12 is a flowchart showing the initial detection position extraction process of the position calibration information collection device according to the first embodiment of the present invention
- FIG. 13 is a block diagram showing a configuration of a position calibration information collecting apparatus according to a modification of the first embodiment of the present invention.
- FIG. 14 is a block diagram showing a configuration of a position calibration information collecting apparatus according to another modification of the first embodiment of the present invention
- FIG. 15 is a block diagram showing the configuration of the position calibration information collecting apparatus according to the second embodiment of the present invention
- FIG. 16A is a diagram showing an example of a last detected position history database when the camera last detected a person in the position calibration information collecting apparatus according to the second embodiment of the present invention;
- FIG. 16B is a diagram showing an example of the last detected position history database when the tag reader last detects a person in the position calibration information collecting apparatus according to the second embodiment of the present invention
- FIG. 17 is a flowchart showing the last detected position acquisition process of the position calibration information collecting apparatus according to the second embodiment of the present invention.
- an observation device for acquiring each as observation information
- First detection position estimation means for estimating a position where the person is first detected in the local coordinate space, based on the feature information of the person, the local coordinates, and the time observed by the observation device
- An entrance / exit position estimation unit that estimates local coordinates of an entrance / exit position of the entrance / exit in the local coordinate space based on the position of the person detected for the first time in the local coordinate space, estimated by the initial detection position estimation unit
- Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means
- Position calibration information calculating means for calculating A position calibration information collecting device is provided.
- a person detection history storing the feature information of the person, the local coordinates of the person, and the time when the person is detected in the observation information respectively acquired by the observation device. Based on the information stored in the database and the person detection history database, the position where the person was first detected in the local coordinate space by the first detection position estimation means was detected as the first detection position information, and the person was detected.
- a first detection position history database that stores the time along with the time, The first detection position estimation means refers to the person detection history database for each of the observation information acquired by the observation device and the person is detected, from the time when the person of the observation information is detected.
- the position calibration information collection device Before another predetermined time, it is determined whether other observation information in which the same feature information as the observation information is stored is stored in the human detection history database, and the same feature information is stored.
- the position where the person is detected in the observation information is estimated as the position where the person is detected for the first time, and the first detection position history
- the position calibration information collection device according to the first aspect, which is stored in a database, is provided.
- the entrance / exit position estimation means refers to the automatic door identification information of the entrance / exit in the vicinity of the position where the person is first detected in the local coordinate space, and
- the position calibration information collecting apparatus according to the first or second aspect is provided, which estimates the local coordinates of the entrance / exit position at.
- the fourth aspect of the present invention further comprises an environment map storage means for storing global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment,
- the position calibration information calculation means uses the global coordinates of the entrance / exit positions of the entrance / exit in the global space stored in the environment map storage means, and uses the global coordinates of the entrance / exit positions and the local coordinates of the entrance / exit positions.
- the position calibration information collection device according to any one of the first to third aspects is provided that calculates the position calibration information for performing calibration related to the position of the observation device based on the above.
- the environment map storage means includes an automatic door identification that identifies whether the door of the entrance / exit in the global space is an entrance / exit that opens automatically or whether the door is an entrance / exit that opens manually.
- the position calibration information calculating unit associates the door of the entrance / exit in the local coordinates with the door of the entrance / exit in the global coordinates based on the automatic door identification information stored in the environment map storage unit, and calculates the position calibration information.
- a position calibration information collecting apparatus according to any one of the first to fourth aspects to be calculated is provided.
- the initial detection position estimating means includes When there are a plurality of local coordinates of the entrance / exit position of the entrance / exit where the door of the entrance / exit is determined to be an entrance / exit automatically opened based on the automatic door identification information, the person is selected for the first time. While determining the position at the coordinate farthest from the center coordinate of the environment among the plurality of detected positions as the local coordinates of the entrance position of the entrance, When there are a plurality of local coordinates of the entrance / exit position of the entrance / exit where the door of the entrance / exit is determined to be an entrance / exit that is manually opened based on the automatic door identification information, the person is selected for the first time.
- a position calibration information collecting device is provided, wherein the center-of-gravity positions of a plurality of detected positions are determined as local coordinates of the entrance / exit position of the entrance / exit.
- a person detection history storing the feature information of the person, the local coordinates of the person, and the time at which the person is detected of the observation information respectively acquired by the observation device. Based on the information stored in the database and the person detection history database, the position where the person was first detected in the local coordinate space by the first detection position estimation means was detected as the first detection position information, and the person was detected.
- a first detection position history database that stores the time along with the time
- the environment map storage means stores information on the position within the global space where the person is likely to stay
- the entrance / exit position estimation means is further configured to determine whether the person is within the local coordinate space based on the person characteristic information stored in the person detection history database, the person's local coordinates, and the time when the person is detected.
- the initial detection position history database stores the position where the person is likely to stay in the local coordinate space by the initial detection position estimation means together with the time when the person was detected
- the entrance / exit position estimation means calculates the local coordinates of the position where the person is likely to stay in the local coordinate space based on the position where the person is likely to stay in the local coordinate space stored in the initial detection position history database.
- the position calibration information calculation means relates to the position of the observation device based on the global coordinates of the position where the person is likely to stay and the local coordinates of the position where the person is likely to stay estimated by the entrance position estimation means.
- a position calibration information collecting apparatus according to a fourth aspect is provided, wherein the position calibration information for performing calibration is calculated.
- the person detection history storing the feature information of the person, the local coordinates of the person, and the time at which the person is detected, of the observation information respectively acquired by the observation device.
- the position where the person was first detected in the local coordinate space by the first detection position estimation means was detected as the first detection position information, and the person was detected.
- a first detection position history database that stores the time along with the time
- the environment map storage means stores information about the position where the person cannot enter
- the entrance / exit position estimation means is further configured to determine whether the person is within the local coordinate space based on the person characteristic information stored in the person detection history database, the person's local coordinates, and the time when the person is detected.
- the initial detection position history database stores the position where the person cannot enter in the local coordinate space by the initial detection position estimation means together with the time when the person was detected
- the entrance / exit position estimation means estimates local coordinates of a position in the local coordinate space where the person cannot enter based on a position where the person cannot enter in the local coordinate space stored in the initial detection position history database.
- the position calibration information calculation means is a calibration related to the position of the observation device based on the global coordinates of the position where the person is likely to stay and the local coordinates of the position where the person cannot enter estimated by the entrance position estimation means.
- the position calibration information collecting device Based on the position calibration information calculated based on the global coordinates of the entrance / exit position of the entrance / exit and the local coordinates of the entrance / exit position installed in the global space by the position calibration information calculating device of the position calibration information collecting apparatus.
- the feature information of the person existing in the environment having the doorway, the local coordinates of the position where the person is detected in the local coordinate space of the environment, and the time when the person is detected And an observation device for acquiring each as observation information, Based on the feature information of the person, the local coordinates, and the time observed by the observation device, a last detected position estimation unit that estimates a position where the person was last detected in the local coordinate space; Entry / exit position estimation means for estimating local coordinates of the entrance / exit position of the entrance / exit in the local coordinate space based on the position where the person was last detected in the local coordinate space, estimated by the end detection position estimation means
- Output position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means
- Position calibration information calculation means for A position calibration information collecting device is provided.
- characteristic information of a person existing in an environment having an entrance / exit local coordinates of a position where the person is detected in the local coordinate space of the environment, and a time when the person is detected are obtained as observation information by the observation device, Based on the feature information of the person, the local coordinates, and the time observed by the observation device, a position at which the person is first detected in the local coordinate space is estimated by an initial detection position estimation unit, Based on the position where the person is first detected in the local coordinate space estimated by the initial detection position estimating means, the local coordinates of the entrance / exit position of the entrance / exit in the local coordinate space are estimated by the entrance / exit position estimating means, Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means Output by calibration information calculation means, A position calibration information collecting method is provided.
- a computer The observation device acquires characteristic information of a person existing in an environment having an entrance, local coordinates of the position where the person is detected in the local coordinate space of the environment, and time when the person is detected as observation information. Function to Based on the feature information of the person, the local coordinates, and the time observed by the observation device, a function of estimating a position at which the person is first detected in the local coordinate space by an initial detection position estimating unit; A function for estimating the local coordinates of the entrance / exit position of the entrance / exit in the local coordinate space by the entrance / exit position estimation unit based on the position where the person is first detected in the local coordinate space, estimated by the initial detection position estimation unit.
- Position calibration information for calibrating the position of the observation device based on the global coordinates of the entrance / exit position of the entrance / exit in the global space of the environment and the local coordinates of the entrance / exit position estimated by the entrance / exit position estimation means
- a position calibration information collection program for realizing the above is provided.
- FIG. 1 is a diagram showing a configuration of a position calibration information collecting apparatus according to the first embodiment of the present invention.
- the position calibration information collection device includes an observation device 101, a human detection history database 102 as an example of a human detection history storage unit, and an initial detection position estimation unit (initial detection position estimation unit) 103.
- An initial detection position history database 104 as an example of an initial detection position history storage unit, an entrance / exit position estimation unit (entrance / entrance position estimation unit) 105, and a calibration parameter as an example of a position calibration information calculation unit that calculates position calibration information
- An acquisition unit (calibration parameter acquisition unit) 106 and an environment map database 107 as an example of an environment map storage unit are provided.
- the entrance / exit position estimation unit 105 and the calibration parameter acquisition unit 106 include configuration information reference units 105a and 106a for reading information from the environment map database 107, respectively.
- FIG. 2 shows a room 201 as a specific example of the living environment.
- the room 201 includes a camera 202 and a UWB (Ultra Wide Band) tag reader system 203 (hereinafter referred to as a tag reader 203).
- the camera 202 and the UWB tag reader system 203 are an example of an observation apparatus 101 that is a component of the position calibration information collection apparatus according to the first embodiment of the present invention.
- a camera 202 and a tag reader 203 are installed near the center of the rectangular ceiling 201d of the rectangular parallelepiped room 201.
- a person 204 exists in the room 201 and enters and leaves the room 201.
- a chair 206 on which the person 204 sits and a desk 207 on which the person 204 cannot enter the arrangement area exist on the floor 201f.
- a door 205A, a door 205B, and a door 205C are installed as an example of an entrance to the room 201 which is a closed environment.
- an arbitrary door among the doors 205A, 205B, and 205C in the room 201 will be representatively described as the door 205.
- a pair of opposing walls 201v and 201w of the room 201 The door 205A and the door 205C are respectively disposed, and the door 205B is disposed on the wall 201b connecting the pair of walls 201a and 201c.
- observation devices two types, the camera 202 and the tag reader 203, are installed in the room 201.
- the present invention can be applied even when only one of the observation devices 101 is installed. .
- FIG. 8 is a flowchart showing an overall process (position calibration information collection process) of the position calibration information collection apparatus.
- the observation apparatus 101 observes the inside of the room 201 and detects a person 204 existing in the room 201 at every predetermined observation period (for example, an arbitrary time of 1 second to 2 seconds or 100 ms).
- the observation apparatus 101 stores the detection result in the human detection history database 102.
- the observation apparatus 101 detects the person 204
- the observation apparatus 101 acquires feature information of the detected person 204 (information indicating characteristics (feature amount) indicating that the person is a person) and local coordinates.
- the observation apparatus 101 stores the detected feature information and local coordinates of the person 204 in the person detection history database 102 together with the time when the person 204 is detected.
- the person 204 and the time when the person 204 is detected are associated with each other and stored in the person detection history database 102.
- the local coordinates are position coordinates (for example, XY coordinates) that are detected by the observation apparatus 101 and indicate the position of the person 204 in the room 201. Therefore, the local coordinates include an arrangement position error of the observation apparatus 101 with respect to global coordinates that are absolute coordinates of the room 201, which will be described later, depending on the installation position of the observation apparatus 101. Therefore, as will be described later, position calibration is required.
- the origin position of the local coordinates can be the pixel at the upper left corner of the image captured by the camera 202.
- the origin position of the local coordinates can be the position of any one of a plurality of base stations connected to the tag reader 203.
- the camera 202 includes an imaging unit 202a that detects a person 204 and an image processing unit 202b that performs image processing on image data.
- an image processing unit 202b provided in the camera 202.
- an image processing method for example, a background subtraction method can be used.
- the image processing unit 202b compares the background image data of the room 201 when the person 204 does not exist and the current image data captured by the camera 202, which have been previously captured by the camera 202 and prepared. Thereafter, an area with different pixel values is extracted as a difference area by the image processing unit 202b.
- the image processing unit 202b can determine that the difference area is sufficiently small with respect to the person 204, the image processing unit 202b It may be determined that the difference area is not the person 204.
- the case where the difference area is sufficiently small with respect to the person 204 may be a case where the number of pixels in the difference area is equal to or less than a threshold set in advance based on the minimum number of pixels that can be recognized as the person 204.
- the detected local coordinates of the person 204 can be set, for example, as the barycentric position of the difference area by the image processing unit 202b.
- step S801 in the flowchart of FIG. That is, in step S801, if the image processing unit 202b determines that the person 204 has been detected using the camera 202, the process proceeds to the next step S802. On the other hand, if the image processing unit 202b determines that the person 204 has not been detected using the camera 202, the processing in step S801 is repeated until the image processing unit 202b determines that the person 204 has been detected. In some cases, although not specifically shown, if the image processing unit 202b determines that the person 204 is not detected, the position calibration information collection process may be terminated. .
- the feature information of the detected person 204 can be, for example, the color distribution of the difference area.
- this is referred to as color feature information.
- the image processing unit 202b determines that the person 204 has been detected using the camera 202. Thereafter, in step S802, the image processing unit 202b extracts the color distribution of the difference area recognized as the person 204 as an example of the feature information of the person 204. Thereafter, the process proceeds to step S803.
- the tag reader 203 includes a tag detection unit 203a that detects a tag, and an information processing unit 203b that calculates the position of the tag based on information detected by the tag detection unit 203a.
- the person 204 In order to detect the person 204 using the tag reader 203, the person 204 is provided with a tag 902 that indicates that the person 204 is present and also has information (tag ID) indicating the person's characteristic information as ID data (identification information). It is necessary to have it in advance.
- tag ID information indicating the person's characteristic information as ID data (identification information). It is necessary to have it in advance.
- FIG. 9 shows an example of detecting the position of the tag 902 by three-point surveying.
- a base station 901A, a base station 901B, and a base station 901C as a tag detection unit 203a connected to the information processing unit 203b are arranged.
- the information processing unit 203b knows local coordinates where the three base stations 901A, 901B, and 901C are arranged.
- the information processing unit 203b has a storage unit in which the local coordinates of the three base stations 901A, 901B, and 901C are stored.
- Each base station 901A, 901B, 901C can calculate the distance to the tag 902 by measuring the time when the radio waves emitted from the base stations 901A, 901B, 901C return from the tag 902. .
- the base station 901A has the tag 902 on an arc whose center is the position of the base station 901A and whose radius is 223.61 cm.
- the base station 901B has the tag 902 on an arc having a radius of 316.23 cm centered on the position of the base station 901B.
- the base station 901C has the tag 902 on an arc having a radius of 141.42 cm centered on the position of the base station 901C.
- the information processing unit 203b can determine that the position where all the three arcs overlap is the position where the tag 902 exists.
- FIG. 9 the tag position detection in the two-dimensional space has been described.
- the arc in FIG. 9 is only a spherical surface, and there is no change in other processes. The above corresponds to the processing in step S801 in the flowchart of FIG.
- step S801 when the tag reader 203 is used and the information processing unit 203b of the tag reader 203 determines that the person 204 is detected (tag ID is detected), the process proceeds to the next step S802.
- the information processing unit 203b of the tag reader 203 detects the person 204.
- the process in step S801 is repeated until it is determined that it has been detected. In some cases, although not specifically shown, if the information processing unit 203b of the tag reader 203 determines that the person 204 is not detected, the position calibration information collection process is terminated. Also good.
- the characteristic information of the detected person 204 can be, for example, ID data (tag ID) stored in the tag 902.
- ID data tag ID
- the process of extracting the feature information that is the tag ID corresponds to the process of step S802 in the flowchart of FIG. Thereafter, the process proceeds to step S803.
- the observation apparatus 101 includes a timer for acquiring information about the time when the person 204 (or the tag ID of the tag 902) is detected.
- observation period between the camera 202 and the tag reader 203 is assumed to be 1 second as an example.
- FIG. 3A shows an example of the human detection history database 102 when the camera 202 detects the human 204.
- FIG. 3B shows an example of the person detection history database 102 when the tag reader 203 detects the person 204.
- the first detection position estimation means 103 estimates the observation ID at which the person 204 is first detected by the observation apparatus 101 from the detection history information of the person 204 stored in the person detection history database 102. Details of the estimation method will be described using the following first detection position history database 104.
- the first detection position history database 104 At least a position where the person 204 is detected for the first time is stored by the first detection position estimation means 103.
- FIG. 4A shows an example of the initial detection position history database 104 when the camera 202 detects the person 204 for the first time
- FIG. 4B shows the case where the tag reader 203 detects the person 204 for the first time.
- the initial detection position estimation unit 103 can store the time when the camera 202 detects the person 204 for the first time, the local coordinates, the color feature information, and the observation ID. It is like that.
- the data of the initial detection position history database 104 in FIG. 4A is created by the initial detection position estimation unit 103 based on the detection history information of the person 204 stored in the human detection history database 102 in FIG. 3A.
- a method of acquiring the first detection position using the data of the first detection position history database 104 of FIG. 4A will be described with reference to the flowchart of FIG.
- step S1201 the first detection position estimation means 103 determines whether or not there is unread data in the human detection history database 102. If the first detection position estimation means 103 determines that there is no unread data, the first detection position acquisition process ends. A method for determining whether or not there is unread data in the first detection position estimation means 103 will be described later.
- Step S1202 is processing when the initial detection position estimation unit 103 determines that unread data exists in step S1201.
- One piece of unread data stored in the person detection history database 102 is read by the initial detection position estimation means 103.
- step S1203 the following processing is performed on the data read in step S1202. That is, the first detection position estimation means 103 determines whether or not data in which the same feature information is stored in the human detection history database 102 between the stored time and N times before is stored. . If it is determined by the first detection position estimation means 103 that the same feature information is stored in the human detection history database 102, the first detection position estimation means 103 is the first detected person Judged not. Then, the process returns to step S1201.
- N can be an observation period of the observation apparatus 101 (for example, an arbitrary time of 1 to 2 seconds or 100 ms). Considering the possibility of the observation apparatus 101 making a detection error of the person 204, a value obtained by multiplying the observation period of the observation apparatus 101 by a constant may be used for N.
- Step S1204 is processing in the case where the initial detection position estimation unit 103 determines that there is data storing the same feature information between the stored time and N times before in step S1203. is there.
- the first detection position estimation means 103 stores the data read in step S1202 in the first detection position history database 104. Thereafter, the process returns to step S1201.
- observation ID CAM_001
- observation ID CAM_005
- a field indicating whether or not the person 204 has been detected for the first time may be provided in the human detection history database 102.
- observation ID CAM_001 is stored because the person whose color feature information is red is detected for the first time.
- a person whose color feature information is red is considered to have left the room 201 at time 2008/09 / 02_12: 00: 08 and entered the room 201 at time 2008/09 / 02_12: 00: 10. It is.
- the observation apparatus 101 here, the camera 202
- the value of N is set to 1 second, which is an example of the observation period of the observation apparatus 101.
- the first detection position estimation unit 103 determines that the person 204 is present in the room 201 when the person 204 having the same color feature information can continuously observe in the observations continuously for each observation period. This is because it is determined that the person 204 has left the room 201 when the person 204 having the same color characteristic information cannot be observed in the continuous observation.
- the camera 202 makes a detection error of the person 204 even though the person 204 exists in the room 201.
- the camera 202 cannot detect the person 204 M times (however, M is an integer greater than 0), even if the initial detection position estimation unit 103 determines that the person 204 has left the room 201. good. That is, when the observation period of the camera 202 is 1 second, if the initial detection position estimation unit 103 determines that the person 204 is not detected for (N ⁇ M) seconds, the initial detection position estimation is that the person 204 has left the room. The determination is made by means 103.
- the red person will not be detected at the time 2008/09 / 02_12: 00: 08, and the red person will be detected the next time after 2 seconds. : 00: 10.
- the first detection position estimation means 103 determines that the person has left the room 201.
- observation ID CAM_001
- observation ID CAM_005
- observation ID CAM_001 and CAM_005 are stored in the first detection position history database 104.
- the first detection position history database 104 of FIG. 4B the time when the tag reader 203 first detects the person 204, the local coordinates, the tag ID, and the observation ID can be stored by the first detection position estimation means 103. It has become. Note that the initial detection position history database 104 in FIG. 4B is created by the initial detection position estimation unit 103 based on the detection history information of the person 204 stored in the human detection history database 102 in FIG. 3B.
- step S1201 the initial detection position estimation means 103 determines whether there is unread data in the human detection history database 102 or not. If the first detection position estimation means 103 determines that there is no unread data, the first detection position acquisition process ends.
- observation ID numbers are assigned to observation IDs in ascending order in the order recorded in the human detection history database 102.
- the first detection position estimation unit 103 stores the read observation ID in the internal memory of the first detection position estimation unit 103 or the like when reading the data recorded in the human detection history database 102. Thereby, the first detection position estimation means 103 can determine that the data next to the observation ID stored in the internal memory is to be read, and if the data does not exist, determine that there is no unread data. Can do.
- Step S1202 is processing when the first detection position estimation unit 103 determines that unread data exists in step S1201, and the first detection position estimation unit 103 reads one piece of unread data stored in the human detection history database 102. .
- step S1203 the following processing is performed on the read data. That is, the first detection position estimation means 103 determines whether or not data in which the same tag ID (feature information) is stored in the human detection history database 102 between the stored time and N times before. Deciding. If the first detection position estimation unit 103 determines that data storing the same tag ID is stored in the person detection history database 102, the first detection position estimation unit 103 is not the first person detected. Determination is made, and the process returns to step S1201.
- N can be an observation period of the observation apparatus 101 (for example, an arbitrary time of 1 to 2 seconds or 100 ms). Considering the possibility of the observation apparatus 101 making a detection error of the person 204, a value obtained by multiplying the observation period of the observation apparatus 101 by a constant may be used for N.
- Step S1204 is processing when the first detection position estimation unit 103 determines in step S1203 that there is no data storing the same tag ID between the stored time and N times before.
- the data read in step S1202 is stored in the first detection position history database 104 by the first detection position estimation means 103 as first detection position information. Thereafter, the process returns to step S1201.
- observation ID TAG_001
- observation ID TAG_005
- observation ID TAG_016 are stored by the first detection position estimation means 103 as observation IDs when the person 204 is detected for the first time. Yes.
- a field indicating whether or not the person 204 is detected for the first time may be provided in the human detection history database 102 without using the initial detection position history database 104.
- observation ID CAM_001 is stored because it is information that a person with a tag ID (feature information) “001” is detected for the first time.
- observation ID CAM_005 is stored because it is information that a person with a tag ID (feature information) “002” is detected for the first time.
- observation ID CAM — 016 is stored because it is information that a person with a tag ID (feature information) “003” is detected for the first time.
- the processing related to the detection error of the observation apparatus 101 is the same as that of the camera 202. Therefore, when it is assumed that the observation apparatus 101 (here, the tag reader 203) does not make a detection error of the person 204, the value of N may be set to 1 second, which is an example of the observation period of the observation apparatus 101. . That is, the first detection position estimation unit 103 determines that the person 204 is present in the room 201 when the person 204 having the same tag ID can continuously observe in the observations continuously for each observation period. This is because it is determined that the person 204 has left the room 201 when the person 204 having the same tag ID cannot be observed in the continuous observation.
- the tag reader 203 makes a detection error of the person 204 even though the person 204 exists in the room 201.
- the initial detection position estimation unit 103 determines that the person 204 has left the room 201. good. That is, when the observation period of the tag reader 203 is 1 second, if the first detection position estimation unit 103 determines that the person 204 is not detected for (N ⁇ M) seconds, the first detection position estimation unit 103 determines that the person 204 It will be judged that it went out of.
- the process of extracting the history information (for example, local coordinates) when the person 204 is detected for the first time by the initial detection position estimation unit 103 is as follows. This corresponds to the process of step S804 in the flowchart of FIG. Thereafter, the process proceeds to step S805.
- step S805 it is determined whether or not the initial detection position estimation means 103 has extracted (existed) the information (for example, local coordinates) of the history of detecting the person 204.
- step S805 when the first detection position estimation unit 103 determines that the history information (for example, local coordinates) from the detection of the person 204 has been extracted, the first detection position estimation unit 103 extracts the history.
- the process of storing the information (for example, local coordinates) in the first detection position history database 104 corresponds to the process of step S806 in the flowchart of FIG. Thereafter, the process proceeds to step S807.
- step S805 if it is determined in step S805 that the initial detection position estimation unit 103 has not extracted the history information (for example, local coordinates) that the person 204 has been detected, the entire process of the position calibration information collection apparatus is terminated. To do.
- the global coordinates are the absolute coordinates of the room 201, unlike the local coordinates, and mean, for example, three-dimensional coordinates with one corner of the floor of the room 201 as the origin.
- the environment map database 107 may be stored in advance in the position calibration information collection device as shown in FIG. Further, instead of storing in advance, as shown in FIG. 14, the same information as the information stored in the environment map database 107 via the Internet 1401 is acquired online by the entrance / exit position estimation unit 105 and the calibration parameter acquisition unit 106, respectively. Also good.
- the environment map database 107 and the Internet 1401 are arranged to be connectable, and necessary information is acquired in advance using the Internet 1401 and stored in the environment map database 107. .
- the updated information may be acquired using the Internet 1401 and stored in the environment map database 107.
- FIG. 5 shows an example of the environment map database 107.
- the environment map database 107 in FIG. 5 stores an environment ID, global coordinates in the furniture room 201 indicated by the environment ID, furniture attributes indicated by the environment ID, and actions that the furniture indicated by the environment ID can take. ing.
- the first one indicates the entrance to the room 201 and corresponds to the door 205.
- they are DOOR_001, DOOR_002, and DOOR_003.
- the second one indicates a position where the person 204 can easily stay in the room 201 (stay position) (for example, it is possible to sit on the furniture), and the chair 206 corresponds to the position.
- environment ID it is CHAIR_004.
- the third one indicates a position where the person 204 cannot enter the room 201 (intrusion is impossible) (for example, the person 204 cannot enter the furniture arrangement area), and the desk 207 corresponds to the third.
- it becomes DESK_005.
- the operation that the furniture indicated by the environment ID can take is stored in advance.
- the door 205 is an automatic door, and it is not necessary for the person 204 to manually open and close the door 205 (the person does not need to manually stop and open / close the door 205C once. ).
- the description will be made with the configuration of the door 205 as described above, the present invention can be operated even when the door 205 is all manually opened or closed.
- a door 205 exists at the entrance and that a person 204 enters and exits the room 202 when the door 205 opens and closes.
- the entrance / exit position estimation means 105 estimates the position of the door 205 (three doors 205A, 205B, 205C) of the room 201 in the local coordinates of the camera 202 and the local coordinates of the tag reader 203, respectively.
- FIG. 6A is an example in which the initial detection position of the person 204 detected by the camera 202 is visualized by the entrance / exit position estimation means 105, and there are 12 positions where the person 204 is detected for the first time. Here, these twelve detection positions (detection position data) need not be positions where the same person is detected.
- the history information of the first detection position in FIG. 6A is a diagram in which the first detection position stored in the first detection position history database 104 is visualized by the entrance / exit position estimation means 105.
- FIG. 6B shows a result of clustering the 12 detection positions in FIG. 6A by the entrance / exit position estimation means 105. It is divided into three clusters of circles (detection position data), triangles (detection position data), and squares (detection position data).
- a cluster classified as a triangle (detection position data) is a cluster 601A
- a cluster classified as a square (detection position data) is a cluster 601B
- a cluster classified as a circle (detection position data) is a cluster 601C.
- an arbitrary cluster among the clusters 601A, 601B, and 601C will be representatively described as the cluster 601.
- a clustering method for example, a k-means method can be used.
- 3 representative positions are randomly selected from 12 positions.
- the three locations represent the number of “entrance / exit” attributes stored in the environment map database 107.
- the data stored in the environment map database 107 is acquired by the configuration information reference unit 105 a included in the entrance / exit position estimation unit 105.
- the distance to each representative position is calculated by the entrance / exit position estimation means 105. Then, the entrance / exit position estimation means 105 determines that the representative position with the shortest distance is the cluster 601 to which the position belongs (assigned).
- each cluster 601 When the assignment of positions to all is completed, the center of gravity of each cluster 601 is set as a new representative point, and the distance to each representative position is calculated by the entrance / exit position estimation means 105 for all positions.
- the entrance / exit position estimation means 105 determines that the representative position with the shortest distance is the cluster 601 to which the position belongs (assigned).
- the creation and assignment of representative points are repeated, and the process ends when there is no change in the cluster 601 to which each position belongs.
- the clustering method is not limited to this.
- the person 204 is first detected by the camera 202 and the tag reader 203 immediately after the door 205 is opened and the room 201 is entered. From this, the position of the door 205 in the local coordinates can be determined by the center-of-gravity position of each position where the initial detection positions of the person 204 are clustered and the entrance / exit position estimation means 105.
- step S808 The process of estimating the local coordinates of the door 205 by the entrance / exit position estimation means 105 based on the information of the initial detection position history of the person 204 stored in the initial detection position history database 104 is the step in the flowchart of FIG. This corresponds to the processing of S807. Thereafter, the process proceeds to step S808.
- the calibration parameter acquisition means 106 as an example of the position calibration information calculation means for calculating the position calibration information includes the position of the door 205 in the local coordinates estimated by the entrance / exit position estimation means 105 and the door 205 stored in the environment map database 107. Based on the position in the global coordinates, parameters as an example of position calibration information necessary for calibration relating to the position of the camera 202 and the position of the tag reader 203 are respectively acquired.
- any of the following five types of parameters can be used as a parameter necessary for position calibration.
- Equation 1 shows the equation for coordinate conversion corresponding to the magnification, parallel movement amount, and rotation angle.
- the environment map database 107 of FIG. 5 there is an item of “operation” for storing the operation of the furniture indicated by the environment ID.
- the environment ID DOOR_003 (door 205C) is “automatic opening / closing” as the information of “operation” that is automatic door identification information, it can be seen that it is an automatic door.
- the door 205 is a manual door, an operation that the person 204 closes the door 205 immediately after the person 204 enters the room 201 occurs.
- the camera 202 and the tag reader 203 detect the person 204 while the person 204 closes the door 205.
- the door 205 is an automatic door
- the person 204 does not need to close the door 205 and continues walking. That is, depending on the observation timing of the camera 202 and the tag reader 203, the person 204 may be detected at the position of the door 205, or may be detected at a position slightly moved from the door 205 to the inside of the room 201. .
- the variation in the detection position at which the initial detection is performed by the automatic door is relatively larger than the variation in the detection position at which the first detection is performed by the manual door (the five circles on the right side of FIG. reference).
- the standard deviation regarding the position of the cluster 601 is obtained by the entrance / exit position estimation means 105, and the standard deviation is the largest.
- the entrance / exit position estimation means 105 may determine that the cluster 601 is the detection position where the initial detection is performed by the automatic door.
- the entrance / exit position estimation means 105 can determine that the cluster 601C is a cluster of the detection position detected for the first time by the door 205C which is an automatic door. If no inversion has occurred in the local coordinates, the cluster 601A has the door 205A based on the positional relationship between the doors 205A and 205B stored in the environment map database 107 and the positional relationship between the clusters 601A and 601B.
- the entrance / exit position estimation unit 105 can determine that the cluster is the detection position detected for the first time in FIG. 5, and the entrance / exit position estimation unit 105 can determine that the cluster 601B is the cluster of the detection position detected for the first time by the door 205B.
- the entrance / exit position estimation means 105 uses the center of gravity of the cluster 601 as the door position.
- the position where the person 204 is detected for the first time varies as described above. Therefore, when the door 205 is an automatic door, the detection position detected at the extreme end in each detection position belonging to the cluster 601C may be determined by the door position 602 and the entrance / exit position estimation means 105 ( (See FIG. 6C).
- the “endmost” here refers to a position at a coordinate farthest from the center coordinate of the room 201 for detecting the position of the entrance / exit arranged on the extension of the walls 201a, 201b, 201c of the room 201. That is.
- the center coordinates of the room 201 in local coordinates are not known before calibration.
- the position of the center of gravity of the positions of all the persons 204 stored in the person detection history database 102 is obtained by the entrance / exit position estimation means 105.
- the detection position that is farthest from the barycentric position among the detection positions belonging to each cluster is determined by the endmost detection position among the detection positions belonging to the cluster and the entrance / exit position estimation means 105.
- the data stored in the environment map database 107 is acquired by the configuration information reference unit 106 a included in the calibration parameter acquisition unit 106.
- the processing for obtaining the parameters necessary for calibration by the calibration parameter acquisition means 106 based on the position of the door 205 in the local coordinates and the position of the door 205 in the global coordinates stored in the environment map database 107 is as described above. This corresponds to the processing in step S808 in the flowchart. Thereafter, the entire processing of the position calibration information collecting apparatus is terminated.
- the observation device 101 always observes the environment
- the initial detection position estimation means 103 always checks the new person detection history in the human detection history database
- the entrance / exit position estimation means 105 newly detects the first initial detection position history database. You may always check your location history.
- the entire position calibration information collecting process is connected to the observation apparatus 101, the human detection history database 102, the initial detection position estimation means 103, the initial detection position history database 104, the entrance / exit position estimation means 105, and the calibration parameter acquisition means 106.
- a control device 1301 for controlling the above may be separately provided. A configuration example to which the control device 1301 is added is shown in FIGS.
- the control device 1301 controls the observation device 101 to observe the environment. Next, when the control device 1301 confirms the presence of new person detection history information in the human detection history database 102, the control device 1301 controls the initial detection position estimation unit 103 to estimate the initial detection position. When the control device 1301 confirms the presence of new initial detection position history information in the initial detection position history database 104, the control device 1301 controls the entrance / exit position estimation means 105 to estimate the entrance / exit position in the local coordinate system. Note that the timing at which the control device 1301 controls the observation device 101 may be determined by an input from the user.
- the number of new person detection history information recorded in the person detection history database 102 necessary for the control device 1301 to start control of the initial detection position estimating means 103 is determined in advance by input from the user. You can keep it.
- the number of pieces of new initial detection position history information recorded in the initial detection position history database 104 necessary for the control device 1301 to start control of the entrance / exit position estimation means 105 is determined in advance by input from the user. You can keep it.
- the global coordinates and the local coordinates of the person staying position are required.
- the person staying position in the global coordinates may be stored in the environment map database 107 in advance.
- environment ID CHAIR_004 is stored as the position where the person 204 stays. This means that, in the furniture in the room 201, a position where the person 204 is likely to stay is determined in advance, and such a position is stored as a position where the person 204 stays in the environment map database 107. To do.
- the person staying position in the local coordinates can be estimated by the entrance / exit position estimation means 105 based on the detection history information of the person 204 stored in the person detection history database 102. This will be described with reference to FIG. 3A.
- the entrance / exit position estimation means 105 may set the coordinates (750, 350) as the person staying position in the local coordinates.
- the determination by the entrance / exit position estimation unit 105 as to whether or not the person 204 is staying is, for example, when the camera 202 detects the person 204 N times at the same position, the entrance / exit position is that the person 204 is staying.
- the estimation unit 105 may make the determination.
- the entrance / exit position estimation means 105 determines that the person 204 is staying. Judgment will be made. Note that the camera 202 or the tag reader 203 may detect the person 204 at a slightly different position every time depending on the observation noise. Therefore, the entrance / exit position estimation unit 105 may determine that the positions detected within ⁇ ⁇ (cm) from the previous detection position are the same position. ⁇ may be, for example, a standard deviation regarding the position error of the camera 202 or the tag reader 203.
- step S1101 the processing in the entrance / exit position estimation means 105 in step S1101 of FIG. Thereafter, the process proceeds to step S1102.
- the global coordinates and the local coordinates of the position where the person cannot enter are required.
- the position where no human intrusion in global coordinates can be stored in the environmental map database 107 in advance.
- environment ID DESK_005 is stored as a position where the person 204 cannot enter. This means that, in the furniture in the room 201, a position where the person 204 cannot enter is determined in advance, and such a position is stored in the environment map database 107 as a position where the person 204 cannot enter. .
- FIG. 7A is an example of history information of the detection position of the person 204 detected by the camera 202.
- the detection position of the person 204 in FIG. 7A is stored as a point having no range.
- the actual person 204 has a certain width (such as a shoulder width). Therefore, the entrance / exit position estimation means 105 performs a process of thickening the detection position of the person 204 in FIG.
- the value of A may match the actual shoulder width of the person 204, or may be set in consideration of the sensor noise level of the observation apparatus 101.
- FIG. 7B is a diagram in which the detection position of the person 204 in FIG. 7A is further thickened by 1 m in each of the XY directions. 7B, it can be seen that an area where no person 204 is detected, an unexplored area 701 is obtained near the center of the floor 201f of the room 201.
- FIG. The position of the center of gravity of the unstepped area 701 may be handled by the entrance / exit position estimation means 105 as a position where a person cannot enter in local coordinates.
- the entrance / exit position estimation means 105 can accurately obtain the position where the person cannot enter in the local coordinates. I can't.
- step S1102 by the entrance / exit position estimation means 105. Thereafter, the process proceeds to step S1103.
- the entrance / exit position estimation means 105 shall perform the calculation of a person stay position and an inaccessible position, respectively.
- step S1103 the calibration parameter acquisition unit 106 acquires the calibration parameters of the camera 202 and / or the tag reader 203 based on (Equation 1).
- the calibration parameter as the calibration information to be obtained is only one combination, and the calculation time in the calibration parameter acquisition unit 106 is shortened to 1/6 of the time required for the previous six combinations.
- step S1103 is the processing of step S1103 by the calibration parameter acquisition means 106.
- the configuration and operation of the position calibration information collection device 110 are as described above.
- the following describes the position calibration apparatus 111 according to the first embodiment of the present invention, which includes the position calibration information collection apparatus 110 and further includes calibration means (calibration unit) 108.
- the configuration and operation of the position calibration information collecting apparatus 110 are the same as those of the position calibration information collecting apparatus 110, and therefore only the calibration means 108 will be described below.
- the position calibration device 111 further includes calibration means 108 in addition to the position calibration information collection device 110, and the calibration of the camera 202 and / or the tag reader 203 using the calibration parameters acquired by the calibration parameter acquisition means 106. Is performed by the calibration means 108.
- the calibration parameter is +5 cm in the X coordinate
- the actual position of the camera 202 or / and the tag reader 203 may be moved by 5 cm in the + X axis direction, or the data acquired by the camera 202 or / and the tag reader 203 may be + X axis. It may be moved 5 cm in the direction.
- FIG. 10 shows a configuration diagram of a position calibration information collecting apparatus provided with the calibration means 108.
- the position where the person 204 is detected for the first time is the position of the door 205 as an entrance to the room 201 which is a closed environment. This makes it possible to automatically collect information necessary for calibration related to the position of the observation apparatus 101 without using a marker or the like prepared in advance in the environment.
- the calibration unit 108 can calibrate the position of the observation apparatus 101 based on the collected calibration information.
- FIG. 15 is a diagram showing a configuration of a position calibration information collecting apparatus according to the second embodiment of the present invention.
- the position calibration information collection device includes an observation device 101, a human detection history database 102 as an example of a human detection history storage unit, and a last detection position estimation unit (last detection position estimation unit). ) 1303, an end detection position history database 1304 as an example of an end detection position history storage unit, an entrance / exit position estimation unit (entrance / entrance position estimation unit) 105, and an example of a position calibration information calculation unit that calculates position calibration information Calibration parameter acquisition means (calibration parameter acquisition unit) 106 and an environment map database 107 as an example of an environment map storage means.
- the means and database other than the last detection position estimation means 1303 and the last detection position history database 1304 are the same as those described in the first embodiment. Further, the last detection position is used instead of the initial detection position for the processing of the entrance / exit position estimation means 105 in the second embodiment.
- the last detection position estimation means 1303 estimates the observation ID at which the person 204 was last detected by the observation apparatus 101 from the detection history information of the person 204 stored in the person detection history database 102. Details of the estimation method will be described using the following last detected position history database 1304.
- FIGS. 16A and 16B show examples of the last detected position history database 1304 when the camera 202 detects the person 204 last and when the tag reader 203 detects the person 204 last.
- the last detection position estimation means 1303 stores the time when the camera 202 last detected the person 204, the local coordinates, the color feature information, and the observation ID. It can be done. Note that the last detection position history database 1304 in FIG. 16A is created by the last detection position estimation means 1303 based on the detection history information of the person 204 stored in the person detection history database 102 in FIG. 3A.
- step S1501 the last detection position estimation means 1303 determines whether or not there is unread data in the human detection history database 102 before the N time from the current time (when the last detection position is acquired). Yes. If the last detection position estimation unit 1303 determines that there is no unread data, the last detection position acquisition process is terminated.
- N can be an observation period of the observation apparatus 101 (for example, an arbitrary time of 1 to 2 seconds or 100 ms). In consideration of the possibility that the observation apparatus 101 makes a detection error of the person 204, a value obtained by multiplying the observation period of the observation apparatus 101 by a constant may be used as N. Further, it is assumed that the last detection position estimation unit 1303 includes a timer.
- Step S1502 is processing when the undetected position estimation unit 1303 determines that unread data exists in step S1501, and one of the unread data stored in the human detection history database 102 is detected once. It is read by the position estimation means 1303.
- step S1503 data in which the same tag ID (feature information) is stored in the data read in step S1502 from the stored time until N times later is stored in the human detection history database 102. It is determined by the last detection position estimation means 1303 whether or not there is. If the last detection position estimation unit 103 determines that the data in which the same tag ID is stored is stored in the person detection history database 102, the last detection position estimation unit 1303 determines that it is not the last detected person. Determination is made, and the process returns to step S1501.
- Step S1504 is processing performed when the last detection position estimation unit 1303 determines in step S1503 that there is no data storing the same tag ID between the stored time and N times later. Yes, the data read in step S1502 is stored in the last detected position history database 1304 by the last detected position estimating means 1303 as last detected position information. Thereafter, the process returns to step S1501.
- the entrance / exit position is estimated by the entrance / exit position estimation means 105.
- the first detection position is used in the first embodiment
- the last detection position is used instead of the first detection position in the second embodiment. Since only the first detection position is replaced with the last detection position and other processes are not changed, a new description is omitted here.
- the position where the person 204 was last detected is the position of the door 205 as an entrance to the room 201 which is a closed environment.
- the calibration unit 108 can calibrate the position of the observation apparatus 101 based on the collected calibration information.
- all of the first detection position and the last detection position may be used as detection positions used by the entrance / exit position estimation means 105.
- the position calibration information collecting apparatus, the position calibration information collecting method, and the position calibration information collecting program according to the present invention can calibrate the position of the observation apparatus without using a marker for position calibration. Therefore, it is particularly useful for a position calibration information collection device including an observation device used in a security system in an office, factory, or home, and a position calibration information collection method and a position calibration information collection program using the observation device. It is.
Abstract
Description
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を推定する初回検出位置推定手段と、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を推定する出入口位置推定手段と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づいて、前記観測装置の位置に関して校正を行うための位置校正情報を算出する位置校正情報算出手段と、
を備えたことを特徴とする位置校正情報収集装置を提供する。
前記位置校正情報収集装置の前記位置校正情報算出装置により前記グローバル空間に設置されている前記出入口の出入口位置のグローバル座標と前記出入口位置のローカル座標とに基づいて算出された前記位置校正情報に基づいて、前記観測装置の位置に関する校正を行う校正手段と、
を備えることを特徴とする位置校正装置を提供する。
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を最後に検出した位置を推定する終回検出位置推定手段と、
前記終回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を最後に検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を推定する出入口位置推定手段と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づいて、前記観測装置の位置に関して校正を行うための位置校正情報を算出する位置校正情報算出手段と、
を備える位置校正情報収集装置を提供する。
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を初回検出位置推定手段で推定し、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を出入口位置推定手段で推定し、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づいて、前記観測装置の位置に関して校正を行うための位置校正情報を位置校正情報算出手段で算出する、
ことを特徴とする位置校正情報収集方法を提供する。
出入口を有する環境内に存在する人の特徴情報と、前記環境のローカル座標空間において前記人が検出された位置のローカル座標と、前記人が検出された時刻とを観測情報として観測装置でそれぞれ取得する機能と、
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を初回検出位置推定手段で推定する機能と、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を出入口位置推定手段で推定する機能と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づいて、前記観測装置の位置に関して校正を行うための位置校正情報を位置校正情報算出手段で算出する機能と、
を実現させるための位置校正情報収集プログラムを提供する。
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を推定する初回検出位置推定手段と、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を推定する出入口位置推定手段と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づいて、前記観測装置の位置に関して校正を行うための位置校正情報を算出する位置校正情報算出手段と、
を備えたことを特徴とする位置校正情報収集装置を提供する。
前記人検出履歴データベースに記憶されている情報を基に、前記初回検出位置推定手段により前記ローカル座標空間内で前記人を初めて検出した位置を、初回検出位置情報として、前記人を検出した時刻と共に記憶する初回検出位置履歴データベースとをさらに備え、
前記初回検出位置推定手段は、前記観測装置で取得されかつ前記人が検出された前記観測情報のそれぞれについて、前記人検出履歴データベースを参照して、前記観測情報の前記人が検出された時刻から所定時刻前までの間に、前記観測情報と同じ特徴情報が記憶されている他の観測情報が、前記人検出履歴データベースに記憶されているか否かを判断し、同じ特徴情報が記憶されている他の観測情報が前記人検出履歴データベースに記憶されていなかった場合には、当該観測情報の前記人が検出された位置を、前記人を初めて検出した位置と推定して、前記初回検出位置履歴データベースに記憶する、第1の態様に記載の位置校正情報収集装置を提供する。
前記位置校正情報算出手段は、前記環境マップ記憶手段に記憶されている前記グローバル空間における前記出入口の前記出入口位置のグローバル座標を利用して、前記出入口位置のグローバル座標と前記出入口位置のローカル座標とに基づいて、前記観測装置の位置に関する校正を行うための前記位置校正情報を算出する、第1~3のいずれか1つの態様に記載の位置校正情報収集装置を提供する。
前記位置校正情報算出手段は、前記環境マップ記憶手段に記憶された前記自動ドア識別情報に基づき、前記ローカル座標における出入口のドアと前記グローバル座標における出入口のドアとを対応付けて前記位置校正情報を算出する第1~4のいずれか1つの態様に記載の位置校正情報収集装置を提供する。
前記自動ドア識別情報に基づき前記出入口の前記ドアが自動で開く出入口であると判断された前記出入口の前記出入口位置のローカル座標が、前記人を初めて検出した位置として複数あるときには、前記人を初めて検出した複数の位置の中で前記環境の中心座標から最も離れた座標にある位置を、前記出入口の前記出入口位置のローカル座標と決定する一方、
前記自動ドア識別情報に基づき前記出入口の前記ドアが手動で開く出入口であると判断された前記出入口の前記出入口位置のローカル座標が、前記人を初めて検出した位置として複数あるときには、前記人を初めて検出した複数の位置の重心位置を、前記出入口の前記出入口位置のローカル座標と決定する、第5の態様に記載の位置校正情報収集装置を提供する。
前記人検出履歴データベースに記憶されている情報を基に、前記初回検出位置推定手段により前記ローカル座標空間内で前記人を初めて検出した位置を、初回検出位置情報として、前記人を検出した時刻と共に記憶する初回検出位置履歴データベースとをさらに備え、
更に前記環境マップ記憶手段には、前記グローバル空間内であってかつ前記人が滞在しやすい位置に関する情報が記憶されており、
前記出入口位置推定手段は、さらに、前記人検出履歴データベースに記憶されている前記人の特徴情報と前記人のローカル座標と前記人を検出した時刻とに基づき、前記ローカル座標空間内で前記人が滞在しやすい位置を推定し、
前記初回検出位置履歴データベースは、前記初回検出位置推定手段により前記ローカル座標空間内で前記人が滞在しやすい位置を、前記人を検出した時刻と共に記憶し、
前記出入口位置推定手段は、前記初回検出位置履歴データベースに記憶された前記ローカル座標空間内で前記人が滞在しやすい位置に基づいて、前記ローカル座標空間における前記人が滞在しやすい位置のローカル座標を推定し、
前記位置校正情報算出手段は、前記人が滞在しやすい位置のグローバル座標と、前記出入口位置推定手段により推定された前記人が滞在しやすい位置のローカル座標とに基づいて、前記観測装置の位置に関する校正を行うための前記位置校正情報を算出することを特徴とする第4の態様に記載の位置校正情報収集装置を提供する。
前記人検出履歴データベースに記憶されている情報を基に、前記初回検出位置推定手段により前記ローカル座標空間内で前記人を初めて検出した位置を、初回検出位置情報として、前記人を検出した時刻と共に記憶する初回検出位置履歴データベースとをさらに備え、
更に前記環境マップ記憶手段には、前記人が侵入できない位置に関する情報が記憶されており、
前記出入口位置推定手段は、さらに、前記人検出履歴データベースに記憶されている前記人の特徴情報と前記人のローカル座標と前記人を検出した時刻とに基づき、前記ローカル座標空間内で前記人が侵入できない位置を推定し、
前記初回検出位置履歴データベースは、前記初回検出位置推定手段により前記ローカル座標空間内で前記人が侵入できない位置を、前記人を検出した時刻と共に記憶し、
前記出入口位置推定手段は、前記初回検出位置履歴データベースに記憶された前記ローカル座標空間内で前記人が侵入できない位置に基づいて、前記ローカル座標空間における前記人が侵入できない位置のローカル座標を推定し、
前記位置校正情報算出手段は、前記人が滞在しやすい位置のグローバル座標と、前記出入口位置推定手段により推定された前記人が侵入できない位置のローカル座標とに基づいて、前記観測装置の位置に関する校正を行うための前記位置校正情報を算出することを特徴とする第4の態様に記載の位置校正情報収集装置を提供する。
前記位置校正情報収集装置の前記位置校正情報算出装置により前記グローバル空間に設置されている前記出入口の出入口位置のグローバル座標と前記出入口位置のローカル座標とに基づいて算出された前記位置校正情報に基づいて、前記観測装置の位置に関する校正を行う校正手段と、
を備えることを特徴とする位置校正装置を提供する。
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を最後に検出した位置を推定する終回検出位置推定手段と、
前記終回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を最後に検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を推定する出入口位置推定手段と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づく、前記観測装置の位置の校正のための位置校正情報を出力する位置校正情報算出手段と、
を備える位置校正情報収集装置を提供する。
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を初回検出位置推定手段で推定し、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を出入口位置推定手段で推定し、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づく、前記観測装置の位置の校正のための位置校正情報を位置校正情報算出手段で出力する、
ことを特徴とする位置校正情報収集方法を提供する。
出入口を有する環境内に存在する人の特徴情報と、前記環境のローカル座標空間において前記人が検出された位置のローカル座標と、前記人が検出された時刻とを観測情報として観測装置でそれぞれ取得する機能と、
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を初回検出位置推定手段で推定する機能と、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を出入口位置推定手段で推定する機能と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づく、前記観測装置の位置の校正のための位置校正情報を位置校正情報算出手段で出力する機能と、
を実現させるための位置校正情報収集プログラムを提供する。
<位置校正情報収集装置の構成の説明>
図1は、本発明の第一実施形態に係る位置校正情報収集装置の構成を示した図である。
にそれぞれドア205Aとドア205Cとが配置され、一対の壁201a,201cをつなぐ壁201bにドア205Bが配置されている。
観測装置101は、所定の観測周期(例えば1秒~2秒の任意の時間又は100ms)毎に、部屋201内を観測し、部屋201内に存在する人204の検出を行う。観測装置101は、検出結果を人検出履歴データベース102に記憶する。観測装置101で人204を検出すると、観測装置101は、検出した人204の特徴情報(人であることを示す特徴(特徴量)を表す情報)とローカル座標とを取得する。そして、観測装置101が、検出した人204の特徴情報とローカル座標を、人204を検出した時刻と共に人検出履歴データベース102に記憶する。これによって、人204と、人204を検出した時刻とが対応付けされて、人検出履歴データベース102に記憶されている。ここで、ローカル座標とは、観測装置101で検出される、部屋201内での人204の位置を示す位置座標(例えば、XY座標)である。従って、ローカル座標には、観測装置101の設置位置に依存し、後述する部屋201の絶対座標であるグローバル座標に対する観測装置101の配置位置誤差が含まれている。そのため、後述するように、位置の校正が必要となる。
人検出履歴データベース102には、観測装置101により人204が検出された時刻と、観測装置101により検出された人204の特徴情報と、人204が観測装置101により検出されたローカル座標とが、観測装置101により記憶される。
初回検出位置推定手段103は、人検出履歴データベース102に記憶されている人204の検出履歴の情報から、人204が観測装置101で初めて検出された観測IDを推定する。推定方法の詳細は、以下の初回検出位置履歴データベース104を用いて説明する。
初回検出位置履歴データベース104には、少なくとも、人204が初めて検出された位置が初回検出位置推定手段103により記憶される。
環境マップデータベース107には、少なくとも、部屋201への出入口となるドア205のグローバル座標が予め記憶されている。ここで、グローバル座標とは、ローカル座標とは異なり、部屋201の絶対座標であり、一例として、部屋201の床の1つの隅を原点とする三次元座標を意味する。
出入口位置推定手段105は、カメラ202のローカル座標とタグリーダ203のローカル座標とにおける部屋201のドア205(3つのドア205A,205B,205C)の位置をそれぞれ推定する。
位置校正情報を算出する位置校正情報算出手段の一例としての校正パラメータ取得手段106は、出入口位置推定手段105で推定されたドア205のローカル座標における位置と、環境マップデータベース107に記憶されたドア205のグローバル座標における位置とに基づいて、カメラ202の位置とタグリーダ203の位置とに関する校正に必要となる、位置校正情報の一例としての、パラメータをそれぞれ取得する。
ドア205は3箇所存在するため、グローバル座標のドア205の位置とローカル座標のドア205の位置の組み合わせは6通り存在する。その6通り全てにおいて、(式1)により展開される連立方程式を解き、上述した5種類のパラメータをそれぞれ求める。ここで、カメラ202とタグリーダ203とは、部屋201の天井201dから床201fを真下に見下ろすように設置されている。つまり、
仮に部屋201内にドア205が1箇所しか存在しない場合、連立方程式が作成できずに前記パラメータが校正パラメータ取得手段106で求められないことになる。このような場合には、人204が滞在しやすい位置(人滞在位置)又は侵入することができない位置に関する情報を出入口位置推定手段105で求めて、校正パラメータ取得手段106で利用することができる。
ステップS1101における、人204の滞在位置(人滞在位置)について説明する。
図7Aは、カメラ202が検出した人204の検出位置の履歴の情報の一例である。図7Aにおける人204の検出位置は、範囲を持たない点として記憶されている。しかし、実際の人204は、ある一定の幅(肩幅など)を持っている。そこで、図7Aの人204の検出位置をXY方向にそれぞれAcmずつ太らせる処理を出入口位置推定手段105で施して黒い丸印として表すこととする。Aの値は、実際の人204の肩幅に合わせても良いし、観測装置101のセンサノイズの大きさを考慮して設定しても構わない。
以下は、前記位置校正情報収集装置110を備えるとともに、さらに、校正手段(校正部)108も備えた、本発明の第一実施形態に係る位置校正装置111について説明する。なお、前記位置校正情報収集装置110の構成及び作用については、前記位置校正情報収集装置110と同一であるため、以下では、校正手段108についてのみ、説明する。
前記位置校正装置111は、前記位置校正情報収集装置110に加えて、更に、校正手段108を備えており、校正パラメータ取得手段106が取得した校正パラメータを用いてカメラ202又は/且つタグリーダ203の校正を校正手段108で行うものである。校正パラメータがX座標に+5cmであった場合、実際のカメラ202又は/且つタグリーダ203の位置を+X軸方向に5cm移動させても良いし、カメラ202又は/且つタグリーダ203が取得したデータを+X軸方向に5cm移動させても良い。
以上のような構成により、人204を初めて検出した位置を閉環境である部屋201内への出入口としてのドア205の位置であると推定できる。これによって、環境内に予め用意したマーカ等を使用することなく、自動で、観測装置101の位置に関する校正に必要な情報収集を行うことができる。また、収集された校正の情報を基に、観測装置101の位置の校正を校正手段108で行うことができる。
<位置校正情報収集装置の構成の説明>
図15は、本発明の第二実施形態に係る位置校正情報収集装置の構成を示した図である。
終回検出位置推定手段1303は、人検出履歴データベース102に記憶されている人204の検出履歴の情報から、人204が観測装置101で最後に検出された観測IDを推定する。推定方法の詳細は、以下の終回検出位置履歴データベース1304を用いて説明する。
終回検出位置履歴データベース1304には、少なくとも、人204が最後に検出された位置が終回検出位置推定手段1303により記憶される。
また、終回検出位置推定手段1303は、タイマを備えているものとする。
Claims (12)
- 出入口を有する環境内に存在する人の特徴情報と、前記環境のローカル座標空間において前記人が検出された位置のローカル座標と、前記人が検出された時刻とを観測情報としてそれぞれ取得する観測装置と、
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を推定する初回検出位置推定手段と、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を推定する出入口位置推定手段と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づく、前記観測装置の位置の校正のための位置校正情報を出力する位置校正情報算出手段と、
を備える位置校正情報収集装置。 - 前記観測装置でそれぞれ取得された前記観測情報の前記人の前記特徴情報と前記人の前記ローカル座標と前記人を検出した前記時刻とを記憶する人検出履歴データベースと
前記人検出履歴データベースに記憶されている情報を基に、前記初回検出位置推定手段により前記ローカル座標空間内で前記人を初めて検出した位置を、初回検出位置情報として、前記人を検出した時刻と共に記憶する初回検出位置履歴データベースとをさらに備え、
前記初回検出位置推定手段は、前記観測装置で取得されかつ前記人が検出された前記観測情報のそれぞれについて、前記人検出履歴データベースを参照して、前記観測情報の前記人が検出された時刻から所定時刻前までの間に、前記観測情報と同じ特徴情報が記憶されている他の観測情報が、前記人検出履歴データベースに記憶されているか否かを判断し、同じ特徴情報が記憶されている他の観測情報が前記人検出履歴データベースに記憶されていなかった場合には、当該観測情報の前記人が検出された位置を、前記人を初めて検出した位置と推定して、前記初回検出位置履歴データベースに記憶する、請求項1に記載の位置校正情報収集装置。 - 前記出入口位置推定手段は、前記ローカル座標空間内で前記人を初めて検出した位置の近傍において、前記出入口の自動ドア識別情報を参照して、前記ローカル座標空間における出入口位置のローカル座標として推定する、請求項1又は2に記載の位置校正情報収集装置。
- 前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標が記憶されている環境マップ記憶手段をさらに備えて、
前記位置校正情報算出手段は、前記環境マップ記憶手段に記憶されている前記グローバル空間における前記出入口の前記出入口位置のグローバル座標を利用して、前記出入口位置のグローバル座標と前記出入口位置のローカル座標とに基づいて、前記観測装置の位置に関する校正を行うための前記位置校正情報を算出する、請求項1又は2に記載の位置校正情報収集装置。 - 前記環境マップ記憶手段には、前記グローバル空間における前記出入口のドアが自動で開く出入口か、又は、ドアが手動で開く出入口であるかを識別する自動ドア識別情報を記憶しており、
前記位置校正情報算出手段は、前記環境マップ記憶手段に記憶された前記自動ドア識別情報に基づき、前記ローカル座標における出入口のドアと前記グローバル座標における出入口のドアとを対応付けて前記位置校正情報を算出する請求項1又は2に記載の位置校正情報収集装置。 - 前記初回検出位置推定手段は、
前記自動ドア識別情報に基づき前記出入口の前記ドアが自動で開く出入口であると判断された前記出入口の前記出入口位置のローカル座標が、前記人を初めて検出した位置として複数あるときには、前記人を初めて検出した複数の位置の中で前記環境の中心座標から最も離れた座標にある位置を、前記出入口の前記出入口位置のローカル座標と決定する一方、
前記自動ドア識別情報に基づき前記出入口の前記ドアが手動で開く出入口であると判断された前記出入口の前記出入口位置のローカル座標が、前記人を初めて検出した位置として複数あるときには、前記人を初めて検出した複数の位置の重心位置を、前記出入口の前記出入口位置のローカル座標と決定する、請求項5に記載の位置校正情報収集装置。 - 前記観測装置でそれぞれ取得された前記観測情報の前記人の前記特徴情報と前記人の前記ローカル座標と前記人を検出した前記時刻とを記憶する人検出履歴データベースと
前記人検出履歴データベースに記憶されている情報を基に、前記初回検出位置推定手段により前記ローカル座標空間内で前記人を初めて検出した位置を、初回検出位置情報として、前記人を検出した時刻と共に記憶する初回検出位置履歴データベースとをさらに備え、
更に前記環境マップ記憶手段には、前記グローバル空間内であってかつ前記人が滞在しやすい位置に関する情報が記憶されており、
前記出入口位置推定手段は、さらに、前記人検出履歴データベースに記憶されている前記人の特徴情報と前記人のローカル座標と前記人を検出した時刻とに基づき、前記ローカル座標空間内で前記人が滞在しやすい位置を推定し、
前記初回検出位置履歴データベースは、前記初回検出位置推定手段により前記ローカル座標空間内で前記人が滞在しやすい位置を、前記人を検出した時刻と共に記憶し、
前記出入口位置推定手段は、前記初回検出位置履歴データベースに記憶された前記ローカル座標空間内で前記人が滞在しやすい位置に基づいて、前記ローカル座標空間における前記人が滞在しやすい位置のローカル座標を推定し、
前記位置校正情報算出手段は、前記人が滞在しやすい位置のグローバル座標と、前記出入口位置推定手段により推定された前記人が滞在しやすい位置のローカル座標とに基づいて、前記観測装置の位置に関する校正を行うための前記位置校正情報を算出する請求項4に記載の位置校正情報収集装置。 - 前記観測装置でそれぞれ取得された前記観測情報の前記人の前記特徴情報と前記人の前記ローカル座標と前記人を検出した前記時刻とを記憶する人検出履歴データベースと
前記人検出履歴データベースに記憶されている情報を基に、前記初回検出位置推定手段により前記ローカル座標空間内で前記人を初めて検出した位置を、初回検出位置情報として、前記人を検出した時刻と共に記憶する初回検出位置履歴データベースとをさらに備え、
更に前記環境マップ記憶手段には、前記人が侵入できない位置に関する情報が記憶されており、
前記出入口位置推定手段は、さらに、前記人検出履歴データベースに記憶されている前記人の特徴情報と前記人のローカル座標と前記人を検出した時刻とに基づき、前記ローカル座標空間内で前記人が侵入できない位置を推定し、
前記初回検出位置履歴データベースは、前記初回検出位置推定手段により前記ローカル座標空間内で前記人が侵入できない位置を、前記人を検出した時刻と共に記憶し、
前記出入口位置推定手段は、前記初回検出位置履歴データベースに記憶された前記ローカル座標空間内で前記人が侵入できない位置に基づいて、前記ローカル座標空間における前記人が侵入できない位置のローカル座標を推定し、
前記位置校正情報算出手段は、前記人が滞在しやすい位置のグローバル座標と、前記出入口位置推定手段により推定された前記人が侵入できない位置のローカル座標とに基づいて、前記観測装置の位置に関する校正を行うための前記位置校正情報を算出する請求項4に記載の位置校正情報収集装置。 - 請求項1又は2に記載の前記位置校正情報収集装置と、
前記位置校正情報収集装置の前記位置校正情報算出装置により前記グローバル空間に設置されている前記出入口の出入口位置のグローバル座標と前記出入口位置のローカル座標とに基づいて算出された前記位置校正情報に基づいて、前記観測装置の位置に関する校正を行う校正手段と、
を備える位置校正装置。 - 出入口を有する環境内に存在する人の特徴情報と、前記環境のローカル座標空間において前記人が検出された位置のローカル座標と、前記人が検出された時刻とを観測情報としてそれぞれ取得する観測装置と、
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を最後に検出した位置を推定する終回検出位置推定手段と、
前記終回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を最後に検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を推定する出入口位置推定手段と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づく、前記観測装置の位置の校正のための位置校正情報を出力する位置校正情報算出手段と、
を備える位置校正情報収集装置。 - 出入口を有する環境内に存在する人の特徴情報と、前記環境のローカル座標空間において前記人が検出された位置のローカル座標と、前記人が検出された時刻とを観測情報として観測装置でそれぞれ取得し、
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を初回検出位置推定手段で推定し、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を出入口位置推定手段で推定し、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づく、前記観測装置の位置の校正のための位置校正情報を位置校正情報算出手段で出力する、
位置校正情報収集方法。 - コンピュータに、
出入口を有する環境内に存在する人の特徴情報と、前記環境のローカル座標空間において前記人が検出された位置のローカル座標と、前記人が検出された時刻とを観測情報として観測装置でそれぞれ取得する機能と、
前記観測装置で観測された、前記人の前記特徴情報と前記ローカル座標と前記時刻とに基づき、前記ローカル座標空間内で前記人を初めて検出した位置を初回検出位置推定手段で推定する機能と、
前記初回検出位置推定手段で推定された、前記ローカル座標空間内で前記人を初めて検出した位置に基づいて、前記ローカル座標空間における前記出入口の出入口位置のローカル座標を出入口位置推定手段で推定する機能と、
前記環境のグローバル空間における前記出入口の出入口位置のグローバル座標と前記出入口位置推定手段により推定された前記出入口位置の前記ローカル座標とに基づいて、前記観測装置の位置に関して校正を行うための位置校正情報を位置校正情報算出手段で算出する機能と、
を実現させるための位置校正情報収集プログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/122,477 US8504317B2 (en) | 2009-09-04 | 2010-09-02 | Position calibration information collecting apparatus, position calibration information collecting method, and position calibration information collecting program |
CN201080003736.4A CN102265112B (zh) | 2009-09-04 | 2010-09-02 | 位置矫正信息收集装置以及位置矫正装置 |
JP2010543326A JP4677060B1 (ja) | 2009-09-04 | 2010-09-02 | 位置校正情報収集装置、位置校正情報収集方法、及び位置校正情報収集プログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2009-204385 | 2009-09-04 | ||
JP2009204385 | 2009-09-04 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2011027557A1 true WO2011027557A1 (ja) | 2011-03-10 |
Family
ID=43649112
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2010/005397 WO2011027557A1 (ja) | 2009-09-04 | 2010-09-02 | 位置校正情報収集装置、位置校正情報収集方法、及び位置校正情報収集プログラム |
Country Status (4)
Country | Link |
---|---|
US (1) | US8504317B2 (ja) |
JP (1) | JP4677060B1 (ja) |
CN (1) | CN102265112B (ja) |
WO (1) | WO2011027557A1 (ja) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2016021091A (ja) * | 2014-07-11 | 2016-02-04 | オムロン株式会社 | 部屋情報推定装置、部屋情報推定方法、および空気調和装置 |
JP6041944B1 (ja) * | 2015-08-18 | 2016-12-14 | 佐藤 創一 | 対象者監視システム |
JP6084742B1 (ja) * | 2016-07-01 | 2017-02-22 | 佐藤 創一 | 対象者監視システム |
WO2020026480A1 (ja) * | 2018-07-31 | 2020-02-06 | 清水建設株式会社 | 位置検出システム及び位置検出方法 |
JP7390604B2 (ja) | 2018-12-28 | 2023-12-04 | パナソニックIpマネジメント株式会社 | 推定方法、推定装置、及び、プログラム |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP7046506B2 (ja) * | 2017-06-12 | 2022-04-04 | キヤノン株式会社 | 情報処理装置、情報処理方法及びプログラム |
KR20210030279A (ko) * | 2018-07-04 | 2021-03-17 | 소니 주식회사 | 정보 처리 장치 및 방법, 그리고 프로그램 |
CN111381231A (zh) * | 2018-12-28 | 2020-07-07 | 松下知识产权经营株式会社 | 估计方法、估计装置以及记录介质 |
CN109996182B (zh) * | 2019-04-19 | 2021-01-19 | 无锡艾森汇智科技有限公司 | 一种基于uwb定位与监控结合的定位方法、装置及系统 |
US20210304577A1 (en) * | 2020-03-30 | 2021-09-30 | Wiser Systems, Inc. | Integrated Camera and Ultra-Wideband Location Devices and Related Systems |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10113343A (ja) * | 1996-07-03 | 1998-05-06 | Hitachi Ltd | 動作及び行動の認識方法及び装置及びシステム |
JP2005024540A (ja) * | 2003-06-30 | 2005-01-27 | Microsoft Corp | 携帯型コンピュータデバイスの所在と動態を割り出すシステムおよび方法 |
JP2005257644A (ja) * | 2004-03-15 | 2005-09-22 | Intelligent Cosmos Research Institute | 位置検出装置、位置検出方法および位置検出プログラム |
JP2007333998A (ja) * | 2006-06-15 | 2007-12-27 | Hitachi Ltd | 地図の自動生成装置 |
Family Cites Families (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE69736622T2 (de) | 1996-07-03 | 2007-09-13 | Hitachi, Ltd. | System zur Bewegungserkennung |
JP3657116B2 (ja) | 1997-05-14 | 2005-06-08 | 株式会社日立国際電気 | 物体追跡方法及び物体追跡装置 |
JP2001004330A (ja) | 1999-04-23 | 2001-01-12 | Matsushita Electric Works Ltd | 画像処理装置及びその画像処理方法 |
US7433493B1 (en) | 2000-09-06 | 2008-10-07 | Hitachi, Ltd. | Abnormal behavior detector |
JP4224260B2 (ja) * | 2002-02-18 | 2009-02-12 | 株式会社トプコン | キャリブレーション装置、方法及び結果診断装置、並びにキャリブレーション用チャート |
JP2003284059A (ja) | 2002-03-27 | 2003-10-03 | Toshiba Lighting & Technology Corp | カメラ画像による移動物体追跡装置およびカメラパラメータのキャリブレーション方法並びに装置 |
JP2004030445A (ja) | 2002-06-27 | 2004-01-29 | National Institute Of Advanced Industrial & Technology | 移動ロボットの自己位置推定方法及びシステム、並びにプログラム |
US6925403B2 (en) * | 2003-09-15 | 2005-08-02 | Eaton Corporation | Method and system for calibrating a sensor |
US7312752B2 (en) * | 2003-10-22 | 2007-12-25 | Awarepoint Corporation | Wireless position location and tracking system |
CN100430686C (zh) * | 2004-05-14 | 2008-11-05 | 佳能株式会社 | 信息处理方法和设备 |
JP4034812B2 (ja) * | 2004-10-14 | 2008-01-16 | 松下電器産業株式会社 | 移動先予測装置および移動先予測方法 |
JP2006311111A (ja) * | 2005-04-27 | 2006-11-09 | Daikin Ind Ltd | 位置検知システムおよび位置検知方法 |
WO2007015548A1 (ja) * | 2005-08-04 | 2007-02-08 | Matsushita Electric Industrial Co., Ltd. | 検索物品推定装置及び方法、並びに、検索物品推定装置用サーバ |
JP4743538B2 (ja) | 2006-11-17 | 2011-08-10 | アイシン精機株式会社 | カメラ校正装置 |
JP4822280B2 (ja) | 2007-03-12 | 2011-11-24 | 三菱電機株式会社 | 移動体行動解析装置 |
AU2008283845A1 (en) * | 2007-08-06 | 2009-02-12 | Trx Systems, Inc. | Locating, tracking, and/or monitoring personnel and/or assets both indoors and outdoors |
JP4675368B2 (ja) | 2007-11-21 | 2011-04-20 | 日本電信電話株式会社 | 物体位置推定装置、物体位置推定方法、物体位置推定プログラムおよびそのプログラムを記録した記録媒体 |
-
2010
- 2010-09-02 US US13/122,477 patent/US8504317B2/en active Active
- 2010-09-02 WO PCT/JP2010/005397 patent/WO2011027557A1/ja active Application Filing
- 2010-09-02 JP JP2010543326A patent/JP4677060B1/ja active Active
- 2010-09-02 CN CN201080003736.4A patent/CN102265112B/zh active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10113343A (ja) * | 1996-07-03 | 1998-05-06 | Hitachi Ltd | 動作及び行動の認識方法及び装置及びシステム |
JP2005024540A (ja) * | 2003-06-30 | 2005-01-27 | Microsoft Corp | 携帯型コンピュータデバイスの所在と動態を割り出すシステムおよび方法 |
JP2005257644A (ja) * | 2004-03-15 | 2005-09-22 | Intelligent Cosmos Research Institute | 位置検出装置、位置検出方法および位置検出プログラム |
JP2007333998A (ja) * | 2006-06-15 | 2007-12-27 | Hitachi Ltd | 地図の自動生成装置 |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2016021091A (ja) * | 2014-07-11 | 2016-02-04 | オムロン株式会社 | 部屋情報推定装置、部屋情報推定方法、および空気調和装置 |
JP6041944B1 (ja) * | 2015-08-18 | 2016-12-14 | 佐藤 創一 | 対象者監視システム |
JP6084742B1 (ja) * | 2016-07-01 | 2017-02-22 | 佐藤 創一 | 対象者監視システム |
JP2018004458A (ja) * | 2016-07-01 | 2018-01-11 | 佐藤 創一 | 対象者監視システム |
WO2020026480A1 (ja) * | 2018-07-31 | 2020-02-06 | 清水建設株式会社 | 位置検出システム及び位置検出方法 |
JP2020020645A (ja) * | 2018-07-31 | 2020-02-06 | 清水建設株式会社 | 位置検出システム及び位置検出方法 |
JP7257752B2 (ja) | 2018-07-31 | 2023-04-14 | 清水建設株式会社 | 位置検出システム |
US11898847B2 (en) | 2018-07-31 | 2024-02-13 | Shimizu Corporation | Position detecting system and position detecting method |
JP7390604B2 (ja) | 2018-12-28 | 2023-12-04 | パナソニックIpマネジメント株式会社 | 推定方法、推定装置、及び、プログラム |
Also Published As
Publication number | Publication date |
---|---|
JPWO2011027557A1 (ja) | 2013-02-04 |
CN102265112A (zh) | 2011-11-30 |
CN102265112B (zh) | 2014-07-02 |
JP4677060B1 (ja) | 2011-04-27 |
US20110184685A1 (en) | 2011-07-28 |
US8504317B2 (en) | 2013-08-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP4677060B1 (ja) | 位置校正情報収集装置、位置校正情報収集方法、及び位置校正情報収集プログラム | |
US20160379074A1 (en) | System and a method for tracking mobile objects using cameras and tag devices | |
US20160377698A1 (en) | System and a method for tracking mobile objects using cameras and tag devices | |
US20180137369A1 (en) | Method and system for automatically managing space related resources | |
KR101292870B1 (ko) | 이동 장치의 로케이션 역학을 결정하는 시스템 및 방법, 이동 장치를 추적하는 것을 용이하게 하는 머신 구현 시스템, 무선 장치 로케이션 시스템, 및 컴퓨터 판독가능 기록 매체 | |
JP5147036B2 (ja) | 位置推定装置、位置推定方法及び位置推定プログラム | |
JP4709943B2 (ja) | 物体位置推定システム、物体位置推定装置、物体位置推定方法、及び物体位置推定プログラム | |
CA2934102A1 (en) | A system and a method for tracking mobile objects using cameras and tag devices | |
Chen et al. | A fall detection system based on infrared array sensors with tracking capability for the elderly at home | |
CA2974031A1 (en) | Method of tracking one or more mobile objects in a site and a system employing same | |
CN102576064A (zh) | 用于标识预定义区域内的兴趣点的方法和装置 | |
CN105528571A (zh) | 区域信息估计装置、区域信息估计方法以及空调装置 | |
Qiu et al. | iFrame: Dynamic indoor map construction through automatic mobile sensing | |
Ye et al. | 6-DOF pose estimation of a robotic navigation aid by tracking visual and geometric features | |
US20170339528A1 (en) | Method and system for calculating a position of a mobile communication device within an environment | |
Hosseinzadeh et al. | Modal identification of building structures using vision-based measurements from multiple interior surveillance cameras | |
US11445107B2 (en) | Supervised setup for control device with imager | |
KR20090113746A (ko) | 물체의 공간적 의미정보를 이용한 로봇의 자기위치 추정 방법 | |
US11281899B2 (en) | Method and system for determining occupancy from images | |
US20190362517A1 (en) | Image database creation device, location and inclination estimation device, and image database creation method | |
Braun et al. | Low-cost indoor localization using cameras–evaluating ambitrack and its applications in ambient assisted living | |
JP6830561B1 (ja) | 情報処理装置、情報処理方法、情報処理システム、情報処理プログラム | |
Miyazaki et al. | Estimation of the number of humans and their movement paths in a room using binary infrared sensors | |
Braun et al. | Marker-free indoor localization and tracking of multiple users in smart environments using a camera-based approach | |
CN116385489B (zh) | 基于rgbd相机的目标跟随方法及系统 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201080003736.4 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010543326 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13122477 Country of ref document: US |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10813508 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 10813508 Country of ref document: EP Kind code of ref document: A1 |