WO2013128852A1 - 動線情報生成システム、動線情報生成方法および動線情報生成プログラム - Google Patents
動線情報生成システム、動線情報生成方法および動線情報生成プログラム Download PDFInfo
- Publication number
- WO2013128852A1 WO2013128852A1 PCT/JP2013/000976 JP2013000976W WO2013128852A1 WO 2013128852 A1 WO2013128852 A1 WO 2013128852A1 JP 2013000976 W JP2013000976 W JP 2013000976W WO 2013128852 A1 WO2013128852 A1 WO 2013128852A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- trajectory
- information
- identification information
- link
- trajectory link
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
-
- 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
- G01S5/0294—Trajectory determination or predictive filtering, e.g. target tracking or Kalman filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
- G06Q10/0833—Tracking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
-
- 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
- G06T2207/30241—Trajectory
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
- G07C9/20—Individual registration on entry or exit involving the use of a pass
- G07C9/28—Individual registration on entry or exit involving the use of a pass the pass enabling tracking or indicating presence
Definitions
- the present invention relates to a flow line information generation system, a flow line information generation method, and a flow line information generation program that generate flow line information including identification information of a moving body detected by different sensors and a locus of the moving body.
- the multi-target tracking device described in Patent Literature 1 outputs a tracking device target position information (elevation angle, azimuth angle, distance), and tracking target identification information and position (elevation angle, azimuth angle). Connected to the target identification sensor device. Then, the multi-target tracking device described in Patent Literature 1 generates a hypothesis group including a target track by connecting the position information output by the radar device. In addition, the multi-target tracking device generates an angle track representing the direction in which the target is moving using information output from the target identification sensor device. Further, the multi-target tracking device selects hypotheses included in the hypothesis group by calculating the similarity between each hypothesis and each angle track. The multi-target tracking device also performs processing for reducing hypotheses by deleting hypotheses with low reliability and integrating similar hypotheses.
- the system described in Patent Document 2 determines a trajectory candidate that is considered to represent a trajectory of an individual mobile object, and generates a hypothesis from a pair of the trajectory candidate and the mobile object identification information. Then, this system calculates the likelihood that the identification information is detected for the trajectory candidates included in each hypothesis, and estimates the maximum likelihood hypothesis based on the likelihood. The trajectory candidates are generated based on the detected trajectory fragments.
- hypotheses hypotheses for associating trajectories with identification information
- combining a candidate group of trajectories of mobile objects and identification information of mobile objects and selecting one hypothesis with the highest likelihood
- the method for determining the association result with the identification information of the mobile object has the following problems.
- the hypothesis representing the true state is rejected by the pruning process or if the hypothesis representing the true state is not generated, the hypothesis represents the true state even if the maximum likelihood hypothesis is specified. That is not to say that a hypothesis representing the true state cannot be obtained.
- the present invention provides the correspondence between the trajectory of the moving object and the identification information of the moving object even when it may occur that the exact correspondence between the trajectory of the moving object and the identification information of the moving object cannot be specified. It is an object of the present invention to provide a flow line information generation system, a flow line information generation method, and a flow line information generation program capable of estimating and quantifying the degree of ambiguity of the correspondence relationship.
- a flow line information generation system is trajectory information representing a trajectory of a moving object or a fragment of the trajectory of the moving object due to an interruption in tracking, and includes trajectory identification information for identifying the trajectory information, A set of trajectory information that can be regarded as representing the trajectory of one mobile object using trajectory information including the position information of the mobile object detected during tracking and a list of detection times of the position information.
- a trajectory link candidate and a trajectory link candidate generating means for generating a trajectory link candidate that is a set of trajectory links that include all trajectory information without duplication, and the trajectory link and mobile object identification information detected in the tracking area; For each trajectory link / identification information pair that is a combination of the above, likelihood calculation means for calculating the likelihood that the trajectory link represents the trajectory of the mobile object identified by the mobile object identification information, and the trajectory link / identification information pair Based on the likelihood calculated in the above, for each trajectory link candidate, identification information association means for associating likely mobile object identification information with each trajectory link included in the trajectory link candidate, and included in each trajectory link candidate Based on the result of the association between each trajectory link and the mobile object identification information, the trend of the number of times the trajectory information and the mobile object identification information are associated with each other is identified for each trajectory information.
- the trajectory represented by the trajectory information is identified by the mobile body identification information for each trajectory information based on the association tendency digitizing means and the value indicating the tendency of the number of times the trajectory information and the mobile body identification information are associated with each other.
- a correlation score calculating means for calculating a correlation score indicating the degree of ambiguity of the correspondence relationship corresponding to the trajectory of the mobile object with respect to each mobile object identification information.
- the flow line information generation method includes trajectory information that represents a trajectory of a moving body or a fragment of the trajectory of the moving body due to a tracking discontinuity, and includes trajectory identification information that identifies the trajectory information,
- a set of trajectory information that can be regarded as representing a trajectory of one mobile object by using trajectory information including position information of the mobile object detected during body tracking and a list of detection times of the position information. Is a combination of the trajectory link and the mobile object identification information detected in the tracking area, generating a trajectory link candidate that is a set of trajectory links that includes all trajectory information without omissions.
- each trajectory link / identification information pair calculate the likelihood that the trajectory link represents the trajectory of the moving body identified by the mobile object identification information, and based on the likelihood calculated for each trajectory link / identification information pair , Locus linkage
- the likely mobile object identification information is associated with each trajectory link included in the trajectory link candidate, and each trajectory link included in each trajectory link candidate is associated with the mobile object identification information.
- the tendency of the number of times that the trajectory information and the mobile object identification information are associated with each other is identified, and the tendency of the number of times that the trajectory information and the mobile object identification information are associated with each other is identified.
- a correlation score indicating the degree of ambiguity of the correspondence relationship that the trajectory represented by the trajectory information corresponds to the trajectory of the mobile object identified by the mobile object identification information The identification information is calculated.
- the flow line information generation program is a trajectory information representing a trajectory of a moving object or a fragment of the trajectory of the moving object due to the interruption of tracking in the computer, and the trajectory identification information for identifying the trajectory information.
- Trajectory link candidate generation processing that generates a trajectory link that is a set of information and a trajectory link candidate that is a set of trajectory links that includes all trajectory information without omissions, and the trajectory link and movement detected in the tracking area
- a likelihood calculation process for calculating the likelihood that the trajectory link represents the trajectory of the mobile object identified by the mobile object identification information
- identification information association processing for associating likely mobile object identification information with each trajectory link included in the trajectory link candidate, individual trajectory The number of times the trajectory information and the mobile object identification information are associated with each trajectory information for each individual mobile object identification information from the result of the association between each trajectory link included in the connection candidate and the mobile object identification information.
- the trajectory represented by the trajectory information is moved for each trajectory information based on the correlation trend digitizing process for digitizing the tendency of the trajectory and the value indicating the tendency of the number of times the trajectory information and the moving body identification information are associated with each other.
- a correlation score calculation process for calculating a correlation score indicating the degree of ambiguity of the correspondence relationship corresponding to the trajectory of the mobile object identified by the body identification information with respect to each mobile object identification information. It is characterized in.
- the correspondence between the trajectory of the moving object and the identification information of the moving object can be determined even if it may occur that the exact correspondence between the trajectory of the moving object and the identification information of the moving object may not be specified. It is possible to estimate and quantify the degree of ambiguity of the correspondence.
- step S3 It is a flowchart which shows the process of step S3 concretely. It is a schematic diagram which shows the example of the production
- step S6 concretely. It is a figure which shows the specific example of the bipartite graph produced by step S603. It is a figure which shows the example which added the starting point and the end point to the bipartite graph shown in FIG. It is a flowchart which shows the process of step S7 concretely. It is a flowchart which shows concretely the process of step S7 in the modification of embodiment of this invention. It is a flowchart which shows the example of process progress of step S700. It is a flowchart which shows the example of process progress of step S700. It is a block diagram which shows the example of the minimum structure of the flow line information generation system of this invention.
- identification information is referred to as ID.
- FIG. 1 is a block diagram showing an example of a flow line information generation system of the present invention.
- the flow line information generation system of the present invention includes a trajectory information input unit 1, an ID information input unit 2, a flow line information generation unit 3, and a flow line information analysis unit 4.
- the flow line information generation system includes, for each ID, a score that represents the degree of ambiguity of the correspondence between the trajectory of the moving object moving within the predetermined tracking area 50 and the ID (identification information) of the moving object.
- the flow line information including the position coordinates of the moving body on the locus and the detection time information is generated.
- This score represents the degree of ambiguity of the correspondence relationship that the trajectory corresponds to the trajectory of the moving object identified by the ID.
- this score represents the reliability of the correspondence between the trajectory and the ID. That is, it can be said that the higher the score, the less ambiguity and the higher the reliability.
- this score is referred to as an ID association score.
- each moving body P may come out of the tracking area 50.
- the kind of the moving body P is not specifically limited, A human, an animal, or a thing may be sufficient.
- the trajectory information input unit 1 is a device that detects the position coordinates of the moving body in the tracking area 50 and inputs the position coordinates and the detection time to the flow line information generation unit 3. In the following description, a case where a two-dimensional coordinate is detected as a position will be described as an example.
- the trajectory information input unit 1 assigns the same trajectory number to each set of detected position coordinates and detection time while one moving body can be tracked without interruption.
- the trajectory information input unit 1 determines a new tracking number that does not overlap with any tracking number assigned so far, when tracking of the moving object is started again. The new tracking number is assigned to the set of position coordinates and detection time detected for the moving object after the tracking is resumed.
- the trajectory information input unit 1 inputs to the flow line information generation unit 3 a set of the position coordinates of the detected moving body and its detection time, and a trajectory number assigned to the set.
- a set of a set of position coordinates to which a common trajectory number is assigned and its detection time, and the trajectory number are combined and described as trajectory information.
- Trajectory information for example, (the locus number 1, (time t 1, x-coordinate, y-coordinate), (time t 2, x-coordinate, y-coordinate), ..., (time t n, x coordinate, y coordinate) ) And the like.
- the trajectory information can be distinguished by the trajectory number, and trajectory information that can be specified by one trajectory number is counted as one trajectory information.
- trajectory information that can be specified by one trajectory number is counted as one trajectory information.
- a plurality of pieces of trajectory information having different trajectory numbers can be obtained. These individual trajectory information can be said to represent individual fragments of the trajectory. Further, if tracking is not interrupted for one moving body, one piece of trajectory information is obtained. Therefore, it can be said that the trajectory information represents a trajectory fragmented due to the interruption of tracking or the entire trajectory related to one moving body.
- the trajectory information input unit 1 may detect the attribute information of the moving object, and input the attribute information to the flow line information generation unit 3 together with the position coordinates, the detection time, and the trajectory number.
- attribute information a color, a magnitude
- the trajectory information includes attribute information for each detection time.
- the trajectory information input unit 1 may be any device that detects position information of a moving body in the tracking area 50, specifies the detection time, and is assigned a trajectory number, and needs to detect an ID unique to each moving body. There is no.
- the trajectory information input unit 1 may be realized by, for example, a moving body tracking system using a camera, a floor pressure sensor, a laser range finder, a human sensor, or a radar. When the trajectory information input unit 1 is realized in such a manner, the moving body does not need to hold a device necessary for detecting the mobile body itself. Further, the trajectory information input unit 1 may detect the position coordinates of the moving body in such a manner that the moving body holds equipment necessary for coordinate detection.
- the trajectory information input unit 1 may be realized by a mobile tracking system using a wireless communication device such as GPS (Global Positioning System) or an ultrasonic receiver.
- the trajectory information input unit 1 is preferably installed so that the entire tracking area 50 can be detected without a blind spot, but a blind spot that may be partially undetectable may be generated. This is because the flow line information generation unit 3 described later can estimate a series of flow lines even when a plurality of pieces of track information having different track numbers are generated due to the interruption of tracking. .
- the ID information input unit 2 is a device that acquires the ID of a moving object in the tracking area 50.
- the ID information input unit 2 does not always detect the ID when the moving object exists, and whether the ID can be detected depends on the position of the moving object in the tracking region 50. For example, the detection probability of an ID of a mobile object existing in a place close to the ID information input unit 2 is high, and the detection probability of an ID of a mobile object existing in a place away from the ID information input unit 2 is low.
- FIG. 1 shows one ID information input unit 2, but a plurality of ID information input units 2 may be installed in the tracking area 50.
- Each ID information input unit 2 is assigned in advance an ID information input unit ID (that is, identification information of the ID information input unit) for uniquely identifying the ID information input unit 2.
- the ID information input unit ID is used to determine which ID information input unit 2 has detected the ID of the moving object.
- ID information input unit ID is represented by a number
- the ID information input unit ID will be referred to as an ID information input unit number.
- the ID information input unit ID may be represented by a number other than the number.
- the ID information input unit 2 inputs the detected moving body ID, the detection time, and the detected position of the ID to the flow line information generation unit 3.
- the ID information input unit 2 inputs the ID information input unit number of the ID information input unit 2 together with the ID information input unit 2 to the flow line information generation unit 3. Further, the tracking area 50 is divided in advance according to the position detection resolution of the sensor to be used. Then, the ID information input unit 2 may input to the flow line information generation unit 3 as information on the detection position, in which region of the regions obtained by dividing the tracking region 50 the ID of the moving object is detected. .
- the ID information input unit 2 may represent the detection position by an area ID that uniquely identifies each area into which the tracking area 50 is divided.
- the ID information input unit 2 may represent a region corresponding to the ID detection position with two-dimensional coordinates.
- the ID information input unit 2 detects the ID.
- the area corresponding to the position may be represented by the ID information input unit number of the ID information input unit 2 itself.
- the position coordinates of the moving body input from the trajectory information input unit 1 to the flow line information generation unit 3 and the ID detection position input from the ID information input unit 2 to the flow line information generation unit 3 are expressed in a common coordinate system.
- the flow line information generation unit 3 holds a conversion table for converting information on the positions. Then, the flow line information generation unit 3 converts one or both of the position coordinates input from the trajectory information input unit 1 and the detection position input from the ID information input unit 2.
- the flow line information generation unit 3 may convert the position coordinates input from the trajectory information input unit 1 into the coordinate system used in the ID information input unit 2. Alternatively, the flow line information generation unit 3 may convert the detection position input from the ID information input unit 2 into the coordinate system used in the trajectory information input unit 1. Further, the flow line information generation unit 3 may convert the position coordinates input from the trajectory information input unit 1 and the detection position input from the ID information input unit 2 into another coordinate system.
- the flow line information generation unit 3 uses the area ID and the vertex position coordinates of each area ID represented by the coordinate system used in the trajectory information input unit 1. What is necessary is just to hold
- the flow line information generation unit 3 is represented by the ID information input unit number and the coordinate system used in the trajectory information input unit 1. It is only necessary to hold a conversion table indicating a correspondence relationship with the installation position of each ID information input unit 2.
- the ID information input unit 2 tries to detect the ID of the moving body and the ID cannot be detected, the ID of the moving body is set to “none” and the time and position are input to the flow line information generating unit 3. May be.
- the ID information input unit 2 does not input any information such as an ID, the flow line information generation unit 3 may determine that no ID of the moving object has been detected at that time.
- the ID information input unit 2 may be any device that can detect an ID unique to a moving object and specify its detection time and detection position.
- an RFID reader may be used as the ID information input unit 2.
- the moving body has an IC (Integrated Circuit) card and the ID of the IC card is used as the ID of the moving body, an IC card reader may be used as the ID information input unit 2.
- the mobile body has a wireless LAN terminal and uses the MAC (Media Access Control) address of the wireless LAN (Local Access Network) terminal as the mobile body ID
- an access point may be used as the ID information input unit 2. .
- a barcode reader may be used as the ID information input unit 2.
- the moving body is a person
- a person's face, fingerprint, vein, or the like may be used as the person ID
- the ID information input unit 2 may be a reading device for these IDs.
- you may use together ID information input part 2 from which a detection target differs like a face authentication apparatus and an RFID reader.
- the detection areas of the ID information input devices 2 may be installed so as to overlap each other. Or you may install so that it may not mutually overlap.
- the detection of the position coordinates of the moving object by the trajectory information input unit 1 and the detection of the ID of the moving object by the ID information input unit 2 are performed at the same time.
- the flow line information generation unit 3 receives the position coordinates input from the trajectory information input unit 1, and the like.
- the ID and the like input from the ID information input unit 2 may be buffered, and the information stored in the buffer at regular intervals may be used.
- time synchronization is not established between the trajectory information input unit 1 and the ID information input unit 2
- information from the trajectory information input unit 1 and information from the ID information input unit 2 are sent to the flow line information generation unit 3.
- the flow line information generation unit 3 may set the same detection time for the input position coordinates and ID.
- the flow line information generation unit 3 uses the information input from the trajectory information input unit 1 and the information input from the ID information input unit 2 to associate the ID for each ID calculated for the trajectory information of the moving object.
- the flow line information including the score and the position coordinates of the moving object and the detection time is generated.
- FIG. 2 is a block diagram illustrating a configuration example of the flow line information generation unit 3.
- a trajectory information input unit 1, an ID information input unit 2, and a flow line information analysis unit 4 are also illustrated.
- the flow line information generation unit 3 includes a trajectory link candidate generation unit 31, an ID association likelihood calculation unit 32, an ID association unit 33, an ID association result aggregation unit 34, and an ID association score calculation unit 35. And an ID candidate output means 36 with a score.
- trajectory link candidate a set of trajectory links that includes all the trajectory information determined from the entire position coordinate, detection time, and trajectory number information input from the trajectory information input unit 1 without any omissions.
- FIG. 3 is a schematic diagram schematically showing a trajectory represented by trajectory information. In FIG. 3, four trajectories corresponding to the four trajectory information of the trajectory numbers 1 to 4 are shown. In the following description, each piece of trajectory information having trajectory numbers 1, 2, 3, and 4 will be referred to as trajectory information 1, 2, 3, and 4, respectively. With reference to FIG. 3, examples of trajectory link and trajectory link candidates will be described.
- the trajectory information 1, 2, 3, 4 can be regarded as representing the trajectory of one moving body. Accordingly, each piece of trajectory information 1, 2, 3, 4 corresponds to an independent trajectory link. Further, the set of the trajectory information 1 and the trajectory information 3 can be regarded as representing a fragmented trajectory set related to one moving body. Accordingly, the combination of the trajectory information 1 and the trajectory information 3 also corresponds to the trajectory link. Similarly, a combination of the locus information 1 and the locus information 4 and a combination of the locus information 2 and the locus information 3 correspond to the locus connection.
- the case where the number of pieces of trajectory information included in the trajectory link is exemplified is one or two.
- the trajectory link including one piece of trajectory information and two pieces of trajectory information are included.
- Trajectory link,..., Trajectory link including n pieces of trajectory information can be considered.
- a trajectory link that is, a set of trajectory information 1, 2, 3, 4) including a maximum of four pieces of trajectory information is conceivable.
- the trajectory link candidate is a set of trajectory links that includes all the trajectory information without omissions. Therefore, for example, it is assumed that the individual trajectory information 1, 2, 3 and 4 are four independent trajectory links, and the trajectory information 1, 2, 3 and 4 are expressed as trajectory links 1, 2, 3 and 4. . In this case, a set of trajectory links 1, 2, 3, and 4 can be set as trajectory link candidates.
- the trajectory link candidate includes trajectory information 1, 2, 3, and 4 without omission and without duplication.
- trajectory information 1 and 3 is assumed to be a single trajectory link and represented as a trajectory link 13.
- the trajectory information 2 and 4 are independent trajectory links and are represented as trajectory links 2 and 4.
- a set of trajectory links 13, 2, and 4 can be set as trajectory link candidates.
- trajectory link candidate a set of two trajectory links 13 and 24 can be a trajectory link candidate.
- a set of trajectory information 1, 2, 3, 4 is one trajectory link.
- this one trajectory link can be a trajectory link candidate.
- the set of trajectory information 1, 2, 3, 4 is one trajectory link, but four fragmented trajectories as shown in FIG. It is hard to think of it as a trajectory related to the body.
- the trajectory link candidate generation unit 31 When the trajectory link candidate generation unit 31 generates a trajectory link, a set of trajectory information corresponding to a predetermined condition is not set as trajectory link.
- the trajectory link candidate includes trajectory information 1, 2, 3, 4 without omission.
- trajectory link set that is not a trajectory link candidate is shown.
- a set of trajectory information 1 and 3 is a trajectory link 13
- a set of trajectory information 1 and 4 is a trajectory link 14.
- the trajectory information 2 is defined as trajectory link 2.
- the combination of the trajectory links 13, 14, and 2 does not correspond to the trajectory link candidate because the trajectory information 1 is included redundantly.
- each piece of trajectory information 1, 2, 3, and 4 is four independent trajectory links 1, 2, 3, and 4, respectively.
- the locus information 4 is not included in the set of the locus links 1, 2, and 3. Therefore, a set of trajectory links including only trajectory links 1, 2, and 3 does not correspond to trajectory link candidates.
- the trajectory link candidate generation means 31 determines individual trajectory information from the entire information of position coordinates, detection time, and trajectory number input from the trajectory information input unit 1, and generates trajectory link and trajectory link candidates using the trajectory information. To do. However, the trajectory link candidate generation unit 31 does not perform trajectory link for a set of trajectory information that satisfies a predetermined condition as described above. Specifically, the trajectory link candidate generation unit 31 does not perform trajectory link for a set of trajectory information corresponding to any one of the first and second conditions described below. In other words, the trajectory link candidate generation unit 31 prohibits a set of trajectory information corresponding to at least one of the first condition and the second condition from being a trajectory link.
- the first condition is that a set of trajectory information includes a plurality of pieces of trajectory information including the same detection time, and the position coordinates of the moving object detected at the same detection time included in the plurality of trajectory information. Is a condition that the distance is equal to or greater than a predetermined threshold. Satisfying the first condition indicates that the same moving object is present at a location separated by a threshold or more at the same time. It is inappropriate to set the trajectory information set corresponding to such a condition as trajectory link, and the trajectory link candidate generation unit 31 does not set the trajectory information set corresponding to the first condition as the trajectory link.
- a position coordinate with the oldest detection time is called a start point
- a position coordinate with the newest detection time is called an end point.
- the second condition is that when each piece of trajectory information included in the set of trajectory information is arranged from the oldest in the order of detection time of the start point, the end time of the end point of the trajectory information is detected, and the time axis is set to a new time side.
- the moving speed of the linear distance from the end point of the previous trajectory information to the start point of the next trajectory information, which is obtained from the detection time and position of the start point of the adjacent next trajectory information, is determined as the maximum movement of the moving object.
- the condition is that the speed is exceeded.
- Satisfying the second condition indicates that the moving body is moving at a speed at which it cannot move. It is inappropriate to set the trajectory information set corresponding to such a condition as trajectory link, and the trajectory link candidate generation unit 31 does not set the trajectory information set corresponding to the second condition as the trajectory link.
- the trajectory link candidate generation unit 31 obtains a distance along the movable route as the distance from the end point of the previous trajectory information to the start point of the next trajectory information, and The moving speed of the moving body may be calculated from the distance.
- the third condition is a condition that, in a plurality of pieces of trajectory information, it is determined that the attribute information included in the different trajectory information is not similar based on the similarity obtained from the attribute information included in the different trajectory information. .
- the trajectory link candidate generation means 31 does not have to set trajectory information corresponding to the third condition as trajectory link. Note that the method for calculating the similarity between the attribute information is not particularly limited. Whether the attribute information is similar may be determined by comparing the degree of similarity with a predetermined threshold value.
- trajectory information is assumed to be a trajectory link.
- the trajectory link candidate generation means 31 sets an upper limit on the number of trajectory link candidates, and when the number of generated trajectory link candidates exceeds the upper limit, the number of trajectory link candidates becomes equal to or less than the upper limit.
- pruning of trajectory link candidates may be performed.
- the pruning standard is not particularly limited.
- the trajectory link candidate generation unit 31 may obtain the sum of the times when the position coordinate detection times overlap among the trajectory information included in the trajectory link candidates, and may exclude the trajectory link candidates from the longest time in order. .
- the trajectory link candidate generation unit 31 may obtain the sum of the discontinuity times between the trajectory information in the trajectory links included in the trajectory link candidates, and may exclude the trajectory link candidates having the long total discontinuity times in order.
- the trajectory link candidate generation means 31 excludes the trajectory link candidates in the descending order of the trajectory links included in the trajectory link candidates (in other words, the descending order of the number of moving objects represented by the trajectory link candidates). Also good.
- the trajectory link candidate generation unit 31 uses the trajectory information in the trajectory link included in the trajectory link candidates as the trajectory information.
- the degree of similarity of the attribute information may be calculated, and the trajectory link candidates including a set of trajectory information having a low degree of similarity may be excluded in order.
- a plurality of criteria may be used in combination as a pruning criterion for trajectory link candidates.
- the ID association likelihood calculating unit 32 uses all the trajectory links generated by the trajectory link candidate generating unit 31 and the IDs of all the moving bodies input from the ID information input unit 2 during a predetermined time. , A set of one trajectory link and one ID (hereinafter referred to as a trajectory link / ID pair) is generated as a whole.
- the trajectory link / ID pair represents the association between the trajectory of the moving object and the ID.
- the ID association likelihood calculation means 32 calculates a likelihood indicating the likelihood of this association for each trajectory link / ID pair. That is, the ID association likelihood calculating means 32 calculates a likelihood representing a trajectory of the moving object whose trajectory link is identified by the ID.
- this likelihood is referred to as an ID association likelihood.
- the ID association likelihood and the ID association score are common in that both represent the degree of appropriateness of association between the trajectory and the ID, but differ in the following points.
- the ID association likelihood is a value calculated for associating the optimum ID with the trajectory link, and is calculated by paying attention to the individual trajectory link.
- the ID association score is obtained after the ID is associated with the trajectory link belonging to each trajectory link candidate generated by the trajectory link candidate generation unit 31 based on the ID correlation likelihood. This is a value calculated based on the tendency of association between the trajectory link and the ID in the link candidate. Further, the ID association likelihood is calculated for the trajectory link, while the ID association score is calculated for the trajectory information.
- the ID association likelihood calculation means 32 holds an ID output probability map in advance.
- the ID output probability map is a map in which the tracking area 50 is divided and the probability that the ID of the moving body is detected by the ID information input unit 2 when the moving body exists for each divided area is defined.
- the ID detection probability defined for each area is referred to as an ID output probability.
- the ID output probability map is created in advance for each ID information input unit 2, and the ID association likelihood calculation means 32 holds each ID output probability map. Each ID output probability map covers the entire area in the tracking area 50.
- a mobile object carries in advance a terminal or tag that transmits an ID and moves in each area in the tracking area 50.
- One ID information input unit 2 tried to detect ID a plurality of times (for example, 10 times) while the moving object was present in one area, and the ID information input unit 2 was able to detect the ID.
- Count the number of times.
- the ID output probability in the area is obtained from the number of times the ID information input unit 2 tries to detect the ID and the number of times the ID can be actually detected.
- an ID output probability map relating to the ID information input unit 2 can be created.
- an ID output probability map may be created for each ID information input unit 2 by the same method.
- the ID association likelihood calculation means 32 corresponds to the ID information input unit 2 that detects the ID of the mobile body included in the locus connection / ID pair when calculating the ID association likelihood for each locus connection / ID pair. Select an ID output probability map. Then, the ID association likelihood calculating means 32 specifies the position coordinates of the moving object at each time when the ID of the moving object included in the trajectory link / ID pair is detected from the trajectory link included in the trajectory link / ID pair. The ID output probability in the area including the position coordinates is determined by the ID output probability map. Then, the ID association likelihood calculating unit 32 multiplies ID output probabilities obtained at each ID detection time. This multiplication result is the ID association likelihood.
- the ID association likelihood calculating means 32 may calculate the sum of logarithmic values of the ID output probabilities obtained at each ID detection time and use the calculation result as the ID association likelihood.
- the ID correlating means 33 is an ID likely to each trajectory link included in the trajectory link candidate (that is, A process of associating a likely mobile ID corresponding to the trajectory represented by the trajectory link is performed.
- the ID association unit 33 is included in the trajectory link candidate by specifying a combination that maximizes the result of integration of the ID association likelihood among the combinations of the trajectory link and each ID included in the trajectory link candidate.
- Each trajectory link is associated with an ID.
- the calculation for obtaining the integrated result of the ID association likelihood is executed in step S606 described later.
- the ID association unit 33 does not associate the same ID with a plurality of trajectory links included in one trajectory link candidate.
- the result of performing the most likely ID association with the trajectory link included in the trajectory link candidate is referred to as an ID correlation candidate.
- the ID association unit 33 associates IDs among the trajectory links included in the trajectory link candidates.
- Information indicating that the corresponding ID does not exist is associated with the trajectory link that has not been performed.
- the character string “unknown” may be used as information indicating that the corresponding ID does not exist, and the ID association unit 33 may associate “unknown” with the trajectory link included in the trajectory link candidate.
- the ID association unit 33 associates IDs with all the link candidates included in the trajectory link candidates. Then, the ID association unit 33 regards an ID that is not associated with a connection candidate as not having a corresponding locus connection, and does not associate it with a connection candidate.
- the ID association unit 33 has an ID corresponding to the trajectory link included in the trajectory link candidate. You may decide not to.
- the ID association result totaling unit 34 targets all ID association candidates obtained by the ID association unit 33, and in each ID association candidate, the trajectory information included in the trajectory link corresponds to which ID and how many times. Aggregate what was attached. If the trajectory link is associated with the ID, the trajectory information included in the trajectory link is also associated with the ID.
- the ID association score calculating unit 35 is configured to calculate each ID for each trajectory information based on the count total result obtained by the ID correlation result totaling unit 34 (that is, the number of times each ID is associated with each trajectory information). Normalize the association count. In the present embodiment, the ID association score calculation means 35 normalizes the value of the number of times obtained for each ID so that the sum is 1 for each trajectory information. In the present embodiment, each value obtained as a result is set as an ID association score. For example, assume that a value of 0.2 is obtained for ID “1” and a value of 0.8 is obtained for ID “2” with respect to certain trajectory information. In this case, the ID association score for the trajectory information and ID “1” is 0.2, and the ID association score for the trajectory information and ID “2” is 0.8.
- the scored ID candidate output means 36 generates a list of individual IDs and ID association scores for the individual IDs for each trajectory information, and outputs the list to the flow line information analysis unit 4.
- the flow line information analysis unit 4 analyzes and visualizes the flow line information by using each ID and ID association score list for each piece of trajectory information generated by the scored ID candidate output means 36.
- the flow line information analysis unit 4 specifies the ID association score of the specified ID for each piece of trajectory information, changes the thickness of the flow line according to the ID association score, and displays each flow line on the display device. It may be displayed above.
- the ID association score represents the degree of ambiguity of association between trajectory information and ID. Therefore, the flow line information analysis unit 4 changes the thickness of the trajectory indicated by the trajectory information according to the ID association score, thereby indicating the degree of ambiguity in the association between the trajectory information and the ID, and the flow line. Information can be visualized. An example of visualizing the flow line information in this way is shown in FIG. FIG.
- the flow line information analysis unit 4 may display a smoothly connected flow line by complementing the discontinuity between the trajectories and continuously changing the thickness of the flow line of the complement portion. In this case, the line thickness of the complemented portion changes smoothly.
- FIG. 4 shows a case where the thickness of the trajectory indicated by the trajectory information is changed according to the ID association score.
- the trajectory information is set according to the ID association score by sharing the thickness of each trajectory.
- the darkness of the locus line indicated by may be changed.
- the flow line information analysis unit 4 may display a darker color for a trajectory with a higher ID association score, and display a lighter color for a trajectory with a lower ID association score.
- the flow line information analysis unit 4 displays the smoothly connected flow line by complementing the discontinuity between the trajectories and continuously changing the color density of the flow line of the complement portion. Also good. In this case, the line density of the complemented portion changes smoothly.
- the analysis and visualization by the flow line information analysis unit 4 is not limited to the above example. Hereinafter, other examples of analysis and visualization will be shown.
- the flow line information analysis unit 4 displays that, for an area where the density of the moving object is equal to or greater than a predetermined value, the set of moving objects in the area corresponds to the set of IDs of the moving objects. May be.
- This area may be the same area as each area in the ID output probability map, or may be an area obtained by dividing the tracking area 50 in a manner according to the user's designation.
- FIG. 5 is an explanatory diagram showing an example of deriving the correspondence between such a set of mobile objects and a set of IDs.
- the example shown in FIG. 5 shows an example in which the tracking area 50 is divided into nine. Of these nine regions, it is assumed that the moving bodies 41, 42, and 43 exist in the region A, and the flow line information analysis unit 4 determines that the density of the moving bodies in the region A is high. In this case, the flow line information analysis unit 4 refers to the ID association score for each ID calculated for the trajectory information of each mobile object in the region A, and for each ID, the sum of the ID association scores. Calculate The flow line information analysis unit 4 does not use an ID association score for a moving object that does not exist in the region A for this calculation. In the area A shown in FIG.
- the flow line information analysis unit 4 associates the three moving bodies 41 to 43 existing in the region A with the top three IDs “a”, “b”, “c” of the sum of the ID association scores, and moves It may be displayed that a set of bodies 41 to 43 is associated with a set of IDs “a”, “b”, and “c”.
- the flow line information analysis unit 4 displays the result of associating the set of the moving object and the ID set in this way. May be.
- the flow line information analysis unit 4 may display the possibility that there is a moving object having an attribute specified by the user for each area.
- the flow line information analysis unit 4 can access a storage device that stores a database in which each mobile object ID is associated with an attribute of the mobile object.
- the flow line information analysis unit 4 determines each area obtained by dividing the tracking area 50 by the size of the area specified by the user, for example.
- the flow line information analyzing unit 4 receives information specifying an attribute from the user. For example, an attribute such as “woman in 40s” is designated. Then, the flow line information analysis unit 4 reads the ID group of the moving object corresponding to the specified attribute from the database, and selects the flow line information that matches the ID group.
- the flow line information analysis unit 4 determines, based on the selected flow line information, which ID moving body exists in which area at which time. Then, the flow line information analysis unit 4 cumulatively adds, for each area, the score of the ID corresponding to the designated attribute among the IDs of the mobile objects existing in the area. It can be said that the region where the addition result is higher is a place where the specified attribute (in this example, a woman in her 40s) is more likely to exist.
- the flow line information analysis unit 4 displays the tracking area 50 by changing the display mode of each area in the tracking area 50 according to the addition result. For example, the flow line information analysis unit 4 may display the tracking area 50 in the heat map format illustrated in FIG.
- Trajectory link candidate generation means 31, ID association likelihood calculation means 32, ID association means 33, ID association result aggregation means 34, ID association score calculation means 35, and scored ID candidate output means 36 Is realized by, for example, a CPU (Central Processing Unit) of a computer that operates according to a flow line information generation program.
- a program storage device (not shown) of the computer stores a flow line information generation program, and the CPU reads the program.
- the trajectory link candidate generation unit 31 and the ID association likelihood calculation unit 32 are read.
- the ID association unit 33, the ID association result aggregation unit 34, the ID association score calculation unit 35, and the scored ID candidate output unit 36 may be operated.
- the trajectory link candidate generation means 31, the ID association likelihood calculation means 32, the ID association means 33, the ID association result totaling means 34, the ID association score calculation means 35, and the scored ID candidate output means 36 are respectively provided. It may be realized by other hardware.
- FIG. 7 is a flowchart illustrating an example of processing progress of the flow line information generation unit 3.
- the trajectory link candidate generation means 31 acquires a set of the trajectory number, the position coordinates of the moving body, and the detection time thereof from the trajectory information input unit 1 (step S1).
- the trajectory link candidate generation means 31 stores a set of trajectory numbers, position coordinates, and detection times that are sequentially input as a data structure having a list of detection times and position coordinates for a predetermined past time for each trajectory number. .
- information on the trajectory number 1 from time t 1 to t n is (trajectory number 1, (time t 1 , x coordinate, y coordinate), (time t 2 , x coordinate, y coordinate),.
- the trajectory link candidate generation unit 31 generates and stores trajectory information including a trajectory number, a list of detection times and detection coordinates for a certain past time. Specifically, when a set of a trajectory number, a position coordinate, and a detection time is newly input, the trajectory link candidate generation unit 31 adds the newly input position coordinate and the list to the list corresponding to the trajectory number. A set of detection times may be added. In addition, the trajectory link candidate generation unit 31 may delete the detection time and position coordinates before a certain past time from the list.
- the trajectory link candidate generation means 31 retains the detection time of the position coordinate to which the trajectory number is first assigned without deleting it.
- this time is referred to as a track number assignment time.
- the trajectory link candidate generation means 31 rearranges the trajectory information group from the current time to the past fixed time in order from the oldest to the newest assigned trajectory number (step S2).
- the trajectory link candidate generation means 31 generates a trajectory link and a trajectory link candidate using the trajectory information group rearranged in step S2 (step S3).
- 8 and 9 are flowcharts specifically showing the process of step S3.
- FIG. 10 is a schematic diagram illustrating an example of a process of generating a trajectory link and a trajectory link candidate.
- step S3 the trajectory link candidate generation means 31 selects the trajectory number of the trajectory information with the oldest trajectory number assignment time from the trajectory information rearranged in step S2 (see FIG. 7), and updates the trajectory information.
- the middle trajectory link and the updating trajectory link candidate are set (step S301).
- the updating trajectory link is a trajectory link in the generation process.
- the updating trajectory link candidate is a trajectory link candidate in the generation process.
- the trajectory link is represented by listing the trajectory numbers in parentheses. In addition, by listing such trajectory links in curly brackets, trajectory link candidates are represented. For example, when the trajectory information of the trajectory number 1 is selected in step S301, the trajectory link candidate generation unit 31 sets the trajectory information as an updating trajectory link (trajectory number 1). In addition, the trajectory link candidate generation unit 31 sets the trajectory information as an updating trajectory link candidate ⁇ (trajectory number 1) ⁇ .
- the trajectory link candidate generation unit 31 determines whether there is a trajectory number not selected in steps S301 and S303 among the trajectory numbers of each trajectory information (step S302). If it is determined that there is no unselected trajectory number (No in step S302), the process in step S3 is terminated.
- the trajectory link candidate generation unit 31 selects the trajectory number of the trajectory information with the oldest trajectory number assignment time from the trajectory information with no trajectory number selected. (Step S303). In the example shown in FIG. 10, the trajectory information of trajectory numbers 1 to n is arranged in order of the trajectory number assignment time. Since trajectory number 1 has already been selected in step S301, trajectory link candidate generation means 31 selects trajectory number 2.
- the trajectory link candidate generation unit 31 adds the trajectory information of the trajectory number selected in Step S303 (here, trajectory number 2) to the updating trajectory link candidate, and updates the updating trajectory link candidate. Perform the process.
- the trajectory link candidate generation means 31 determines whether there is an updating trajectory link candidate that has not been selected in step S305 among the updating trajectory link candidates generated at the end of step S303 (step S303). S304).
- the trajectory link candidate generation unit 31 selects one of the unselected updating trajectory link candidates (step S305).
- the above-described updating trajectory link candidate ⁇ (trajectory number 1) ⁇ will be described as an example.
- the trajectory link candidate generation unit 31 determines whether there is an updating trajectory link that is not selected in step S307 among the updating trajectory links included in the updating trajectory link candidate selected in step S305. (Step S306). If there is an updating trajectory link that has not yet been selected in step S307 in the selected updating trajectory link candidates (Yes in step S306), the trajectory link candidate generation means 31 is not selected from the updating trajectory link candidates. Update trajectory link is selected (step S307). In this example, (trajectory number 1) is selected from the updating trajectory link candidates ⁇ (trajectory number 1) ⁇ .
- the trajectory link candidate generation unit 31 can link the trajectory information of the trajectory number selected in step S303 to the updating trajectory link selected in step S307 (in other words, whether it can be added). ) Is determined (step S308).
- the trajectory link candidate generation unit 31 determines that the set of trajectory information in the updating trajectory link selected in step S307 and the trajectory information of the trajectory number selected in step S303 is the first condition and the second It may be determined whether or not connection is possible by determining whether or not the condition is met. That is, the trajectory link candidate generation unit 31 selects the trajectory number selected in step S303 for the selected updating trajectory link if the set of trajectory information does not correspond to either the first condition or the second condition. It may be determined that the trajectory information can be linked.
- the trajectory link candidate generation means 31 selects the updated trajectory link in step S303. What is necessary is just to determine with the locus
- each of the trajectory information of the trajectory number 1 and the trajectory information of the trajectory number 2 includes the position coordinates of the moving object detected at the same detection time, and the distance between the position coordinates is If it is equal to or greater than the threshold value, the first condition is met.
- the trajectory link candidate generation unit 31 determines that the trajectory information of the trajectory number 2 cannot be linked to the updating trajectory link (trajectory number 1).
- the trajectory link candidate generation means 31 calculates the trajectory information of the trajectory number 1 from the detection time and position of the end point of the trajectory information of the trajectory number 1 and the detection time and position of the start point of the trajectory information of the trajectory number 2. The speed at which the moving body has moved the distance from the end point to the start point of the trajectory information of the trajectory number 2 is calculated. If this speed exceeds the predetermined maximum moving speed of the moving body, the set of trajectory information of trajectory numbers 1 and 2 corresponds to the second condition. In this case, the trajectory link candidate generation unit 31 determines that the trajectory information of the trajectory number 2 cannot be linked to the updating trajectory link (trajectory number 1).
- the trajectory link candidate generation unit 31 includes the trajectory information set described above. It may be determined whether or not the third condition is satisfied. When the third condition is satisfied, the trajectory link candidate generation unit 31 may determine that the trajectory information of the trajectory number selected in step S303 cannot be linked to the updating trajectory link.
- step S307 If the first and second conditions are met and it is determined in step S307 that connection is not possible (No in step S308), the trajectory link candidate generation unit 31 proceeds to step S306 again.
- the trajectory link candidate generation unit 31 updates the trajectory link in the trajectory link candidate selected in step S305.
- a new updating trajectory link candidate is generated by connecting the updating trajectory link selected in step S307 to the trajectory information of the trajectory number selected in step S303 (step S309).
- the trajectory link candidate generation means 31 is a new updating trajectory link candidate ⁇ added with the trajectory information of the trajectory number 2 to (trajectory number 1) in the updating trajectory link candidate ⁇ (trajectory number 1) ⁇ . (Trajectory number 1, trajectory number 2) ⁇ is generated.
- step S309 the trajectory link candidate generation unit 31 proceeds to step S306 again.
- the trajectory link candidate generation unit 31 again does not select the updating trajectory link included in the updating trajectory link candidates selected in step S305, in step S307. It is determined whether or not there is an updating trajectory link (step S306).
- the trajectory link candidate generation means 31 has already selected (trajectory number 1) in the updating trajectory link candidate ⁇ (trajectory number 1) ⁇ selected in step S305, and other unselected updating trajectories. It is determined that there is no connection.
- the trajectory link candidate generation unit 31 sets the trajectory information of the trajectory number selected in step S303 as a new updating trajectory link.
- the updating trajectory link is added to the updating trajectory link candidate selected in step S305 (step S310).
- the trajectory link candidate generation unit 31 selects the trajectory information of the trajectory number 2 in step S303, and sets this trajectory information as a new updating trajectory link (trajectory number 2). Further, the trajectory link candidate generation means 31 selects the updating trajectory link candidate ⁇ (trajectory number 1) ⁇ in step S305.
- the trajectory link candidate generation unit 31 adds the updating trajectory link candidate (trajectory number 2) to the updating trajectory link candidate ⁇ (trajectory number 1) ⁇ , and the updating trajectory link candidate ⁇ (trajectory number 1), (trajectory number). Number 2) ⁇ .
- the trajectory link candidate generation means 31 creates a new updating trajectory link candidate ⁇ ( Locus number 1, locus number 2) ⁇ is generated. Further, the trajectory link candidate generation means 31 sets ⁇ (trajectory number 1) ⁇ to ⁇ (trajectory number 1), (trajectory number 2) ⁇ in step S310. Accordingly, ⁇ (trajectory number 1), (trajectory number 2) ⁇ is generated at least in step S310 based on ⁇ (trajectory number 1) ⁇ , and when the process proceeds to step S309, ⁇ (trajectory number 1, trajectory Number 2) ⁇ is also generated. As described above, when the process proceeds to step S309, the number of updating trajectory link candidates increases. In step S310, the updating trajectory link candidate selected in step S305 is updated.
- updating trajectory link candidate generated in step S309 and the updating trajectory link candidate updated in step S310 are all assumed to be unselected as to whether or not the selection in step S305 has been performed. It becomes. Further, each updating trajectory link included in these updating trajectory link candidates is also initialized as unselected with respect to whether or not the selection in step S307 has been performed.
- an upper limit of the number of trajectory link candidates may be set.
- the trajectory link candidate generation unit 31 updates the updating trajectory link so that the number of updating trajectory link candidates is equal to or less than the upper limit.
- Candidates may be pruned.
- the pruning standard is not particularly limited.
- the trajectory link candidate generation unit 31 obtains, for example, the sum of the times when the position coordinate detection times overlap between the trajectory information included in the updating trajectory link candidates, and the updating trajectory link having a long time. You may exclude in order from a candidate.
- the trajectory link candidate generation means 31 obtains the sum of the discontinuity times between the trajectory information in each of the updating trajectory links included in the updating trajectory link candidates, and sequentially excludes from the updating trajectory link candidates having the long total discontinuity times. May be. Further, the trajectory link candidate generation unit 31 may exclude the trajectory link candidates in the descending order of the update trajectory link included in the update trajectory link candidates. Further, when the trajectory information includes the attribute information of the moving object, the trajectory link candidate generation unit 31 uses the attribute information in the trajectory information between the trajectory information in the updating trajectory link included in the updating trajectory link candidate. The degree of similarity may be calculated, and the updating trajectory link candidates including a set of trajectory information having a low degree of similarity may be excluded in order.
- the trajectory link candidate generation unit 31 proceeds to step S304 again and performs the operations after S304.
- the updating trajectory link candidate generated at the end of step S303 is only ⁇ (trajectory number 1) ⁇ , and ⁇ (trajectory number 1) ⁇ has already been selected. 31 determines that there is no unselected updating trajectory link candidate (No in step S304).
- step S303 If there is no unselected updating trajectory link candidate that has been generated at the end of step S303 (No in step S304), the trajectory link candidate generation unit 31 proceeds to step S302. And the locus
- step S302 two updating trajectory link candidates ⁇ (trajectory number 1, trajectory number 2) ⁇ , ⁇ (trajectory number 1), (trajectory number 2) ⁇ as in the above example. Is generated. Then, it is assumed that the trajectory link candidate generation unit 31 determines that there is an unselected trajectory number in step S302 and selects trajectory number 3.
- the trajectory link candidate generation means 31 determines that there is an unselected updating trajectory link candidate in step S304, and selects ⁇ (trajectory number 1, trajectory number 2) ⁇ in step S305, for example. In this case, the trajectory link candidate generation unit 31 selects (trajectory number 1, trajectory number 2) in step S307 and then proceeds to step S309. As a new updating trajectory link candidate, ⁇ (trajectory number 1, trajectory Number 2, trajectory number 3) ⁇ .
- the trajectory link candidate generation unit 31 updates ⁇ (trajectory number 1, trajectory number 2) ⁇ to ⁇ (trajectory number 1, trajectory number 2), (trajectory number 3) ⁇ .
- the trajectory link candidate generation means 31 determines again that there is an unselected updating trajectory link candidate in step S304, and selects ⁇ (trajectory number 1), (trajectory number 2) ⁇ in step S305. In this case, the trajectory link candidate generation means 31 selects (trajectory number 1) in step S307 and then proceeds to step S309. As a new updating trajectory link candidate, ⁇ (trajectory number 1, trajectory number 3), (Trajectory number 2) ⁇ is generated. Further, the trajectory link candidate generation means 31 selects (trajectory number 2) in step S307 and then moves to step S309. As a new updating trajectory link candidate, ⁇ (trajectory number 1), (trajectory number 2, Trajectory number 3) ⁇ is generated.
- the trajectory link candidate generation unit 31 changes ⁇ (trajectory number 1), (trajectory number 2) ⁇ to ⁇ (trajectory number 1), (trajectory number 2), (trajectory number 3) ⁇ . Update to
- step S304 If it is determined in step S304 that there is no unselected updating trajectory link candidate, the trajectory link candidate generation unit 31 proceeds to step S302 again. If the trajectory link candidate generation unit 31 determines in step S302 that there is no unselected trajectory number (No in step S302), the trajectory link candidate generation unit 31 ends the process in step S3. Each updating trajectory link candidate obtained at this time is determined as a trajectory link candidate. Further, each updating trajectory link included in each trajectory link candidate is determined as a trajectory link.
- the ID association likelihood calculating means 32 acquires an ID for uniquely identifying the moving object, an ID detection time, and an ID detection position from the ID information input unit 2 (step S4). Since the ID information input unit 2 also inputs the ID information input unit number of the ID information input unit 2 itself to the flow line information generation unit 3, the ID association likelihood calculating means 32 also determines the ID information input unit number. Acquire together.
- FIG. 11 shows an example of each area obtained by dividing the tracking area 50 into a lattice shape. In this example, the case where the ID association likelihood calculating unit 32 acquires the area IDs of these areas from the ID information input unit 2 as detection positions will be described as an example.
- ID information is distinguished from “ID” included in the ID information.
- ID information is a set of an ID, a detection time, and a detection position of a mobile object. Further, “ID” included in the ID information is the ID of the mobile object itself.
- the ID association likelihood calculating means 32 stores ID information for a certain past time.
- the length of the “fixed time” is the same as the length of the “fixed time” when the list of detection times and position coordinates for the past fixed time is included in the trajectory information.
- the ID association likelihood calculation means 32 uses the trajectory link generated in step S3 and the ID of the moving body included in each ID information acquired in step S4 to generate a trajectory link / ID pair. Generate everything. Then, the ID association likelihood calculating means 32 calculates an ID association likelihood for each trajectory link / ID pair (step S5).
- ID association likelihood calculation a specific example of ID association likelihood calculation will be shown.
- FIG. 11 a case where two ID information input units 2a and 2b are arranged is taken as an example. Further, it is assumed that the ID association likelihood calculating means 32 stores the ID information illustrated in FIG. In this example, it is assumed that the ID information input unit 2a detects an ID at times t1, t2, and t3, and the ID information input unit 2b detects an ID at times t9, t10, and t11.
- the ID association likelihood calculating means 32 stores in advance an ID output probability map corresponding to the ID information input unit 2a and an ID output probability map corresponding to the ID information input unit 2b.
- FIG. 13 is a schematic diagram of an ID output probability map corresponding to the ID information input unit 2a. In the ID output probability map shown in FIG. 13, the ID output probability of the area 12 where the ID information input unit 2a exists is 0.8, and the ID output probability of the eight surrounding areas is 0.5. The ID output probability of other areas is 0.1.
- FIG. 14 is a schematic diagram of an ID output probability map corresponding to the ID information input unit 2b. In the ID output probability map shown in FIG. 14, the ID output probability of the area 58 where the ID information input unit 2 b exists is 0.8, and the ID output probability of the eight surrounding areas is 0.5. The ID output probability of other areas is 0.1.
- FIG. 15 and FIG. 16 are schematic views schematically showing the trajectory link on the ID output probability map. 15 and FIG. 16, the trajectory connection including the trajectory information of the trajectory numbers 1 and 2 is shown. Hereinafter, this trajectory link is referred to as a trajectory link A.
- FIG. 15 is a schematic diagram in which the trajectory link A is superimposed on the ID output probability map corresponding to the ID information input unit 2a.
- FIG. 16 is a schematic diagram in which the trajectory link A is superimposed on the ID output probability map corresponding to the ID information input unit 2b.
- FIG. 17 and FIG. 18 are schematic diagrams schematically showing the trajectory link on the ID output probability map.
- 17 and 18 show the trajectory link including the trajectory information of the trajectory numbers 3 and 4.
- this trajectory link is referred to as a trajectory link B.
- FIG. 17 is a schematic diagram in which the trajectory link B is superimposed on the ID output probability map corresponding to the ID information input unit 2a.
- FIG. 18 is a schematic diagram in which the trajectory link B is superimposed on the ID output probability map corresponding to the ID information input unit 2b.
- the ID association likelihood calculating unit 32 calculates each position of the moving body at each detection time of the IDs constituting the trajectory link / ID pair. Then, it is specified based on the trajectory link constituting the trajectory link / ID pair. Then, the ID association likelihood calculating means 32 specifies the ID output probability at each position from the ID output probability map of the ID information input unit 2 that detects the ID. Further, the ID association likelihood calculating means 32 calculates the product of the ID output probabilities and sets the calculation result as the ID association likelihood. The ID association likelihood calculating means 32 may calculate the sum of the logarithmic values of the identified ID output probabilities and use the calculation result as the ID association likelihood.
- the process of calculating the ID association likelihood of the trajectory link / ID pair of the trajectory link A and ID1 is shown.
- ID1 is detected in the area 12 at times t1, t2, and t3. This ID1 is detected by the moving body input unit 2a (see FIG. 11).
- the ID association likelihood calculating means 32 specifies the position of the moving body from the trajectory link A at the detection time t1 of ID1. As shown in FIG. 15, the position coordinates of the moving object at time t ⁇ b> 1 exist in area 02.
- the ID association likelihood calculating means 32 specifies the area 02 as the position of the moving object at time t1 (see FIG. 15).
- the ID association likelihood calculating means 32 acquires the ID output probability 0.5 of the area 02 from the ID output probability map of the mobile input unit 2a that detected ID1 at time t1. Similarly, the ID association likelihood calculating unit 32 specifies the position of the moving object at the detection time t2 of ID1 based on the trajectory link A. In this case, the ID association likelihood calculating unit 32 specifies the area 12. Then, the ID association likelihood calculating means 32 acquires the ID output probability 0.8 of the area 12 from the ID output probability map of the mobile input unit 2a that detected ID1 at time t2 (see FIG. 15).
- the ID association likelihood calculating means 32 specifies the area 22 as the position of the moving object at the detection time t3 of ID1, and the ID detection probability 0.5 of the area 22 from the ID output probability map of the moving object input unit 2a. (See FIG. 15).
- ID1 is detected in the area 58 at times t9, t10, and t11.
- This ID1 is detected by the moving body input unit 2b (see FIG. 11).
- the ID association likelihood calculating means 32 specifies the area 57 based on the trajectory link A as the position of the moving body at the detection time t9 of ID1, and the ID output probability map of the moving body input unit 2b that detected ID1 at the time t9.
- the ID output probability 0.5 of area 57 is acquired from (see FIG. 16).
- the ID association likelihood calculating unit 32 acquires ID detection probabilities 0 and 8 for the detection time t10 of ID1, and acquires an ID detection probability of 0.5 for the detection time t11 of ID1 (see FIG. 16).
- the ID association likelihood calculating means 32 calculates the product of the ID output probabilities obtained for the detection times t1, t2, t3, t9, t10, and t11 of ID1, and obtains 0.04. This value is the ID association likelihood of the pair of trajectory link A and ID1.
- ID2 is detected in the area 58 at times t9, t10, and t11. This ID2 is detected by the moving body input unit 2b (see FIG. 11).
- the ID association likelihood calculating means 32 specifies the area 57 as the position of the moving object at the detection time t9 of ID2, and detects the ID of the area 57 from the ID output probability map of the moving object input unit 2b. A probability of 0.5 is acquired (see FIG. 16).
- the ID association likelihood calculating means 32 acquires an ID output probability of 0.8 with respect to the detection time t10 of ID2, and acquires an ID detection probability of 0.5 with respect to the detection time t11 of ID2 (see FIG. 16). Then, the ID association likelihood calculating means 32 calculates the product of the ID output probabilities obtained for the detection times t9, t10, and t11 of ID2, and obtains 0.2. This value is the ID association likelihood of the pair of trajectory link A and ID2.
- the ID association likelihood calculating unit 32 specifies the area 15 as the position of the moving object at the detection time t1 of ID1. Then, the ID association likelihood calculation means 32 acquires the ID output probability 0.1 of the area 15 from the ID output probability map of the mobile input unit 2a that detected ID1 at time t1 (see FIG. 17). Similarly, the ID association likelihood calculation means 32 specifies the area 25 as the position of the moving object at the detection time t2 of ID1, and the ID detection probability 0.1 of the area 25 from the ID output probability map of the moving object input unit 2a. (See FIG. 17).
- the ID association likelihood calculating means 32 specifies the area 36 as the position of the moving object at the detection time t3 of ID1, and sets the ID detection probability 0.1 of the area 36 from the ID output probability map of the moving object input unit 2a. Obtain (see FIG. 17).
- the ID association likelihood calculating means 32 specifies the area 48 based on the trajectory link B as the position of the moving object at the detection time t9 of ID1, and the ID output of the moving object input unit 2b that detected ID1 at the time t9.
- the ID output probability 0.5 of area 48 is acquired from the probability map (see FIG. 18).
- the ID association likelihood calculating unit 32 acquires ID detection probabilities 0 and 8 for the detection time t10 of ID1, and acquires an ID detection probability of 0.5 for the detection time t11 of ID1 (see FIG. 18).
- the ID association likelihood calculating means 32 calculates the product of the ID output probabilities obtained for the detection times t1, t2, t3, t9, t10, and t11 of ID1, and obtains 0.0002. This value is the ID association likelihood of the pair of trajectory link B and ID1.
- the ID association likelihood calculating means 32 Based on the trajectory link B, the ID association likelihood calculating means 32 identifies the area 48 as the position of the moving object at the detection time t9 of ID2, and detects the ID of the area 48 from the ID output probability map of the moving object input unit 2b. A probability of 0.5 is acquired (see FIG. 18). Similarly, the ID association likelihood calculating unit 32 acquires the ID output probability 0.8 with respect to the detection time t10 of ID2, and acquires the ID detection probability 0.5 with respect to the detection time t11 of ID2 (see FIG. 18).
- the ID association likelihood calculating means 32 calculates the product of the ID output probabilities obtained for the detection times t9, t10, and t11 of ID2, and obtains 0.2. This value is the ID association likelihood of the pair of trajectory link B and ID2.
- the ID association likelihood calculating means 32 calculates the ID association likelihood of all the trajectory link / ID pairs by the same procedure. When the position coordinates of the moving object at the detection time of the ID of the moving object are not detected, the ID association likelihood calculating unit 32 performs the ID detection time by interpolating or extrapolating the trajectory information in the trajectory link. What is necessary is just to identify the position of the mobile body in and to obtain the ID output probability at that position.
- step S5 the ID association unit 33 associates the most likely ID to each trajectory link included in the trajectory link candidate based on the ID link likelihood for each trajectory link / ID pair calculated in step S5 ( Step S6).
- FIG. 19 is a flowchart specifically showing the process of step S6.
- step S6 the ID association unit 33 determines whether there is a trajectory link candidate not selected in step S602 from the trajectory link candidate group generated in step S3 (step S601). If there is an unselected trajectory link candidate (Yes in step S601), the ID association unit 33 selects the trajectory link candidate that has not been selected in step S602 so far from the trajectory link candidate group generated in step S3. Is selected (step S602).
- the ID association unit 33 creates a bipartite graph in which each trajectory link included in the trajectory link candidate selected in step S602 is a vertex and each ID of the moving object is a vertex (step S603).
- the ID association unit 33 creates a bipartite graph with no edges between the vertices corresponding to the trajectory link and no edges between the vertices corresponding to the ID.
- FIG. 20 is a diagram illustrating a specific example of the bipartite graph created in step S603.
- the example shown in FIG. 20 shows a bipartite graph with the trajectory links 100 and 101 as vertices and ID1 and ID2 as vertices. From the bipartite graph illustrated in FIG. 20, the correspondence between the trajectory links 100 and 101 and ID1 and ID2 is obtained.
- the ID association unit 33 adds a node serving as a start point and a node serving as an end point to both ends of the created bipartite graph (step S604).
- the ID association unit 33 may add a start point connected to each vertex corresponding to the trajectory link and an end point connected to each vertex corresponding to the ID to the bipartite graph.
- FIG. 21 shows an example in which a start point and an end point are added to the bipartite graph shown in FIG. In FIG. 21, the values shown for each side are the costs set in step S605 described below.
- the ID association unit 33 sets a cost for each side of the graph obtained in step S604 (step S605).
- the cost is a value indicating the ease with which the nodes are associated with each other. The smaller the cost is, the more easily the nodes are associated with each other.
- the ID association means 33 calculates the reciprocal of the ID association likelihood of the trajectory link / ID pair of the trajectory link and ID pair for the side connecting the vertex corresponding to the trajectory link and the vertex corresponding to the ID. What is necessary is just to set as cost. For example, when the ID association likelihood of the pair of the trajectory link 100 and ID1 is 0.025, the ID association means 33 is the reciprocal of 0.025 on the side connecting the trajectory link 100 and ID1. 40 is set as the cost. Further, the ID association unit 33 sets the cost of the side connecting the start point and each vertex corresponding to the trajectory link to 0. Similarly, the ID association unit 33 sets the cost of the side connecting the end point and each vertex corresponding to the ID to 0.
- the ID association unit 33 sets the capacity of each side to “1”.
- the capacity of each side is a value indicating how many paths can pass through the side when a route from the start point to the end point is determined. Therefore, for example, in the example shown in FIG. 21, after the route passing through the nodes in the order of “start point, trajectory link 100, ID1, end point” is determined, the capacity of each side between the nodes becomes zero. In this case, the path between the trajectory link 100 and ID1 in the route indicates the correspondence between the trajectory link and ID for one moving body. Further, when searching for a route on a graph corresponding to another moving body, an edge having a capacity of 0 cannot be searched.
- the ID association unit 33 applies the minimum cost flow problem in the graph theory to the graph obtained by the processing up to step S605, and obtains the correspondence between the trajectory link and the ID (step S606).
- the ID association unit 33 sets the flow rate that should satisfy the smaller one of the number of trajectory links and the number of IDs. Then, the ID association unit 33 applies the minimum cost flow problem in the graph theory to the graph obtained by the processing up to step S605, and selects a path with the smallest cost from the start point to the end point.
- the ID association unit 33 sets the flow rate to “2”. Then, the ID association unit 33 includes 2 routes that pass through the nodes in the order of “start point, trajectory link 100, ID1, end point” and 2 routes that add nodes in the order of “start point, trajectory link 101, ID2, end point”. Is selected as a path that satisfies the flow rate “2” with the minimum sum of costs.
- the combination of the trajectory link and the ID included in the route selected in step S606 is the result (ID association candidate) of the most likely ID associated with each trajectory link in the trajectory link candidate selected in step S602. .
- the ID association unit 33 After performing the processing up to step S606 for the trajectory link candidate selected in step S602, the ID association unit 33 repeats the processing after step S601 and executes steps S602 to S606 for each trajectory link candidate. If it is determined that there are no unselected locus connection candidates (No in step S601), the ID association unit 33 ends the process in step S6.
- the ID association result totaling unit 34 associates the trajectory information in the trajectory link included in the ID association candidate with all the ID association candidates generated in step S6. (Step S7).
- the ID association result totaling unit 34 identifies the trajectory information by the trajectory number.
- FIG. 22 is a flowchart specifically showing the process of step S7.
- the ID association result totaling unit 34 identifies a row by the trajectory number of each trajectory information obtained during a certain time, and identifies a column by the ID of the moving object detected at the certain time. Is created (step S701).
- the aggregation map is a table in which a trajectory number is assigned to each row and a moving body ID is assigned to each column, and a portion corresponding to the intersection of the row and the column is referred to as a cell.
- a cell can store a value. The initial value of each cell is zero.
- the ID association result totaling unit 34 determines whether or not there is a trajectory link candidate that has not been selected in step S703 among the trajectory link candidates (step S702). If an unselected trajectory link candidate exists (Yes in step S702), the ID association result totaling unit 34 selects one trajectory link candidate that has not been selected in step S702 so far from each trajectory link candidate. Are selected (step S703).
- the ID association result totaling unit 34 determines whether there is a trajectory link that is not selected in step S705 among the trajectory links included in the trajectory link candidate selected in step S703 (step S704). . If there is a trajectory link that has not been selected in step S705 in the trajectory link candidate (Yes in step S704), the ID association result totaling unit 34 selects an unselected trajectory link from the trajectory link candidates. . Furthermore, the ID association result totaling unit 34 specifies the ID of the moving body associated with the trajectory link (the trajectory link selected from the trajectory link candidates currently focused on) by the ID correlating unit 33 (step S705). ).
- the ID association result totaling unit 34 determines whether or not the trajectory information not selected in step S707 is present among the trajectory information included in the trajectory link selected in step S705 (step S706). .
- the ID association result totaling unit 34 selects the unselected trajectory information from the trajectory link. Further, the ID association result totaling unit 34 adds 1 to the cell value determined by the trajectory number of the trajectory information and the ID of the moving object specified in step S705 in the totalization map. That is, the ID association result totaling unit 34 adds 1 to the value of the cell at the intersection of the row corresponding to the trajectory number and the column corresponding to the ID (step S707).
- the ID association result totaling unit 34 repeats the operations after step S706.
- the ID association result totaling unit 34 repeats the operations in and after step S704.
- the ID association result totaling unit 34 repeats the operations after step S702.
- step S703 when there is no longer any trajectory link candidate that has not been selected in step S703 (No in step S702), the ID association result totaling unit 34 ends the process in step S7.
- the value of each cell obtained after the end of step S7 is obtained by quantifying the tendency of the number of times the trajectory information and the ID of the moving object are associated with each other.
- the ID association score calculation means 35 calculates an ID association score for each ID for each trajectory number based on the result (ie, the value of each cell of the aggregation map) obtained in step S7. Calculate (step S8). For example, the ID association score calculation unit 35 may normalize the value of each cell so that the sum of the values of each cell becomes 1 for each column indicated by each trajectory number in the aggregation map. That is, the ID association score calculation unit 35 may divide the value of each cell in the column by the sum of the values of the cells in the column for each column indicated by each trajectory number.
- the value of each cell obtained as a result is an ID association score relating to a set of a trajectory number and a moving body ID corresponding to the cell. Since the trajectory number represents trajectory information, the ID association score of each ID is obtained for each trajectory information based on the result of step S8.
- the trajectory information, the information associated with the ID association score obtained for each ID of the trajectory information and the list of each ID includes information on the position coordinates of the moving object, the detection time thereof, and the ID of the moving object. It is. Therefore, it can be said that this information is flow line information including an ID association score.
- the ID association score calculation means 35 generates this flow line information.
- the scored ID candidate output means 36 outputs the flow line information including the ID association score obtained by the processing up to step S8 to the flow line information analysis unit 4 (step S9).
- the information output mode from the scored ID candidate output means 36 to the flow line information analysis unit 4 in step S9 is not particularly limited.
- the scored ID candidate output unit 36 may output related information to the flow line information analysis unit 4 at each detection time with reference to the time.
- the scored ID candidate output means 36 includes the detection time t1, the position coordinates of the moving object detected at the time t1, the locus number assigned by the locus information input unit 1 to the time and position coordinates,
- the ID association score for each ID calculated with respect to the trajectory number and a list of each ID may be output to the flow line information analysis unit 4. The same applies to other detection times.
- the scored ID candidate output means 36 may output related information to the flow line information analysis unit 4 for each position coordinate on the basis of the position coordinate of the moving object. For example, the scored ID candidate output means 36 calculates the position coordinate a, the detection time of the position coordinate a, the track number assigned by the track information input unit 1 to the time and the position coordinate, and the track number. The ID association score for each ID and the list of each ID may be output to the flow line information analysis unit 4. The same applies to other position coordinates.
- the scored ID candidate output means 36 may output related information to the flow line information analysis unit 4 for each trajectory number with reference to the trajectory number.
- the scored ID candidate output means 36 analyzes the trajectory information indicated by the trajectory number for each trajectory number, the ID association score for each ID calculated with respect to the trajectory number, and the list of each ID, by the flow line information analysis. You may output to the part 4.
- the scored ID candidate output means 36 may output only the trajectory number, the ID association score for each ID calculated for the trajectory number, and the list of each ID to the flow line information analysis unit 4. .
- the flow line information analysis unit 4 directly acquires the information from the trajectory information input unit 1 and the ID information input unit 2. Also good.
- the analysis mode performed by the flow line information analysis unit 4 using the information acquired in step S9 is not particularly limited.
- the trajectory link candidate generation means 31 generates a trajectory link and a trajectory link candidate. Then, the ID association likelihood calculating means 32 calculates the ID association likelihood for each trajectory link / ID pair. The ID association unit 33 associates a likely ID with the trajectory link in the trajectory link candidate for each trajectory link candidate using the ID correlation likelihood. Therefore, there is a case where a detection error or misdetection of the position coordinates of the moving object in the trajectory information input unit 1 occurs, or a detection error or misdetection of the ID of the moving object occurs in the ID information input unit 2. However, it is possible to generate a plurality of trajectory link candidates associated with IDs that are likely to be related to the trajectory link.
- a hypothesis representing a true state may not be included, but a plurality of hypotheses that are likely to some extent can be generated.
- the ID association result totaling means 34 totalizes which ID the trajectory information in the trajectory link is associated with, and the ID correlation result totaling means 34 corresponds to which ID the trajectory information corresponds to based on the totalization result. Scoring the trend as to what is attached. Therefore, the degree of ambiguity of the correspondence relationship between the individual trajectory information and the ID of each mobile object can be quantified from the tendency of the hypothesis set that is likely to some extent.
- step S707 the value to be added to the cell in step S707 (see FIG. 22) is set to 1.
- the ID association result totaling unit 34 calculates a score representing the likelihood of the trajectory link candidate for each trajectory link candidate, and values in the cells of the total map according to the score. Is added.
- the process of step S7 differs from said embodiment. The processes other than step S7 are the same as those in the above embodiment, and the description is omitted.
- the first score calculation method is to calculate the ID association likelihood regarding the combination of trajectory link and ID obtained as a result of the ID association unit 33 associating the ID with respect to the trajectory link included in the trajectory link candidate.
- This is a method of making a score representing the likelihood of a trajectory link candidate using a product of ID association likelihoods for each trajectory link included in the trajectory link candidate.
- the ID association likelihood is specified for each trajectory link included in the trajectory link candidates, the product of each ID association likelihood is obtained for each trajectory link candidate, and the ID association likelihood for each trajectory link candidate.
- the degree integration result is normalized so that the sum is 1.
- the trajectory information input unit 1 detects not only the position coordinates but also the attribute information at the detection time of the position coordinates of the moving object, and the trajectory link candidate generation unit 31 includes each of the trajectory information in the trajectory information.
- the attribute information of the moving body at the detection time is included. Note that this attribute information is a value quantified numerically. For example, in the case of a moving body color, R, G, and B can be expressed numerically within a range of 0 to 255. Information that can be quantified in this way is used as attribute information.
- the ID association result totaling unit 34 calculates the variance of the attribute information included in each piece of trajectory information in the trajectory link for each trajectory link included in the trajectory link candidate.
- the ID association result totaling unit 34 also calculates a detection time length for each trajectory link included in the trajectory link candidates.
- the detection time length of the trajectory link is the trajectory with the latest detection time of the end point of the trajectory information included in the trajectory link from the detection time of the start point in the trajectory information included in the trajectory link. This is the time until the end point detection time in the information.
- the ID association result totaling unit 34 calculates, for each trajectory link candidate, the average value ( ⁇ ave ) of the variance values of the attribute information calculated for each trajectory link included in the trajectory link candidate and the detection time length.
- the ID association result totaling unit 34 calculates the following equation (1) using ⁇ ave after normalization and T sum after normalization for each trajectory link candidate, and the result is calculated as a trajectory link candidate. A score S representing the likelihood is assumed.
- FIG. 23 is a flowchart specifically showing the process of step S7 (see FIG. 7) in the modification of the embodiment described above.
- the same processes as those in FIG. 22 are denoted by the same reference numerals as those in FIG.
- the ID association result totaling unit 34 calculates a score representing the likelihood of the trajectory link candidate for each trajectory link candidate before the total map creation process (step S701) (step S700).
- a method for calculating a score representing the likelihood of a trajectory link candidate there are the first calculation method and the second calculation method described above. Specific processing progress in these calculation methods will be described later.
- step S700 the ID association result totaling unit 34 executes the processing after step S701.
- the processing in steps S701 to S706 is the same as that in the above-described embodiment (see FIG. 22), and a description thereof will be omitted.
- step S706 when it is determined that there is unselected trajectory information among the trajectory information included in the trajectory link selected in step S705 (Yes in step S706), the ID association result totaling unit 34 has not yet detected the trajectory link. Select trajectory information for selection. Further, the ID association result totaling unit 34 sets the score of the trajectory link candidate selected in step S703 (to the cell value determined by the trajectory number of the trajectory information and the ID of the moving body specified in step S705 in the total map. A score representing the likelihood of the trajectory link candidate) is added (step S707a).
- step S700 a specific process of a score calculation process indicating the likelihood of the trajectory link candidate will be described.
- FIG. 24 is a flowchart illustrating an example of the processing progress of step S700 when the first calculation method is employed.
- step S700 the same processing as in steps S702 to S705 (see FIGS. 22 and 23) already described is performed.
- steps S702 to S705 see FIGS. 22 and 23
- FIG. 24 these processes are denoted by reference numerals S7002 to S7005.
- step S700 the ID association result totaling unit 34 first acquires the ID association likelihood of each trajectory link / ID pair calculated by the ID association likelihood calculating unit 32 (step S7001). After S7001, the ID association result totaling unit 34 executes the processing after step S7002.
- the processing in steps S7002 to S7005 is the same as that in steps S702 to S705 (see FIGS. 22 and 23), and the description thereof is omitted.
- step S7005 the ID association result totaling unit 34 selects a trajectory link and specifies an ID associated with the trajectory link. This process is the same as step S705 already described.
- step S7005 the ID association result totaling unit 34 calculates the ID association likelihood regarding the combination of the trajectory link selected in step S7005 and the ID specified in step S7005 as the score of the trajectory link candidate selected in step S7003 ( The score representing the likelihood of the trajectory link candidate is multiplied (step S7007).
- the ID association result totaling unit 34 initializes the score of each trajectory link candidate (score indicating the likelihood of the trajectory link candidate) to “1” at the start of step S700. Therefore, after selecting a trajectory link candidate in step S7003, when the process first proceeds to step S7007, the ID association result totaling unit 34 multiplies the initial score value “1” by the ID association likelihood.
- the ID association result totaling unit 34 repeats the processing after step S7004. With respect to the trajectory link candidate selected in step S7003, the ID association result totaling unit 34 repeats the processes of steps S7004 to S7007 until there is no unselected trajectory link, so that the ID association result totaling unit 34 stores the ID for each trajectory link included in the trajectory link candidate.
- the result of integration of the association likelihood can be obtained as a score.
- the ID association result totaling unit 34 repeats the processing from step S7002 onward. By repeating the processing from step S7002 until there is no unselected locus connection candidate, the ID association result totaling unit 34 can obtain a score for each locus connection candidate.
- the ID association result totaling unit 34 calculates, for each trajectory link candidate, the score obtained by integrating the ID association likelihoods in step S7007. Normalization is performed so that the sum of the scores becomes 1 (step S70008). That is, the ID association result totaling unit 34 may divide the individual trajectory link candidate scores by the sum of the trajectory link candidate scores. As a result, the value obtained for each trajectory link candidate is a score representing the likelihood of the trajectory link candidate.
- Ending step S7008 ends step S700. Thereafter, the ID association result totaling unit 34 may perform the processing after step S701 shown in FIG.
- a logarithmic value of the ID association likelihood may be used instead of the ID association likelihood.
- the ID association result totaling unit 34 may calculate the logarithmic value of the ID association likelihood of each trajectory link / ID pair in step S7001.
- the ID association result totaling unit 34 may add the logarithmic value of the ID association likelihood to the score of the selected trajectory link candidate.
- FIG. 25 is a flowchart illustrating an example of the processing progress of step S700 when the second calculation method is employed.
- step S700 the same processing as in steps S702 to S704 (see FIGS. 22 and 23) already described is performed. In FIG. 25, these processes are denoted by reference numerals S7002 to S7004.
- step S700 the ID association result totaling unit 34 starts processing from step S7002.
- the processing in steps S7002 to S7004 is the same as that in steps S702 to S704 (see FIGS. 22 and 23), and a description thereof will be omitted.
- step S7004 If it is determined in step S7004 that there is an unselected trajectory link among the selected trajectory link candidates (Yes in step S7004), the ID association result totaling unit 34 selects an unselected trajectory link from the trajectory link candidates. Select. Then, the ID association result totaling unit 34 calculates a variance value of the attribute information included in each piece of trajectory information in the selected trajectory link (step S7011).
- the attribute information is a value obtained by quantifying the attribute, and a variance value of the attribute information can be calculated.
- the ID association result totaling unit 34 may refer to each piece of trajectory information in the selected trajectory link and calculate a variance value of the attribute information detected at each detection time.
- the ID association result totaling unit 34 calculates the detection time length of the trajectory link selected in step S7011 (step S7012).
- the ID association result totaling means 34 has the earliest detection time of the start point among the trajectory information included in the trajectory link from the end point detection time in the trajectory information whose trajectory information included in the trajectory link has the latest end point detection time.
- the detection time length may be calculated by subtracting the detection time of the start point in the trajectory information.
- the ID association result totaling unit 34 repeats the processing after step S7004.
- the ID association result totaling unit 34 repeats the processing from step S7002 onward.
- the ID association result totaling unit 34 calculates, for each trajectory link candidate, an average value of variance values of attribute information related to each trajectory link in the trajectory link candidate. Calculate (step S7013).
- the average value of the variance values obtained for one trajectory link candidate is denoted as ⁇ ave .
- the ID association result totaling unit 34 calculates, for each trajectory link candidate, the total sum of detection time lengths of each trajectory link in the trajectory link candidate (step S7014).
- the sum of the detection time lengths obtained for one trajectory link candidate is denoted as T sum .
- steps S7013 and S7014 are shown.
- description will be made using two trajectory link candidates (suppose U and V).
- the trajectory link candidate U includes two trajectory links U 1 and U 2 . Then, the variance value and detection time length of the attribute calculated for the trajectory link U 1 are ⁇ U1 and T U1 , respectively, and the variance value and detection time length of the attribute calculated for the trajectory link U 2 are ⁇ U2 and T U2 , respectively. And
- the trajectory link candidate V includes three trajectory links V 1 , V 2 , and V 3 .
- the variance value and the detection time length of the attribute calculated for the trajectory link V 1 are set as ⁇ V1 and T V1 , respectively.
- the variance value of the attribute and the detection time length calculated for the trajectory link V 2 are ⁇ V2 and T V2 , respectively.
- the attribute variance value and detection time length calculated for the trajectory link V 3 are ⁇ V3 and T V3 , respectively.
- the ID association result totaling unit 34 may calculate the average value of ⁇ U1 and ⁇ U2 as ⁇ ave (average value of attribute information dispersion values) in the trajectory link candidate U in step S7013. Further, the ID association result totaling unit 34 may calculate an average value of ⁇ V1 , ⁇ V2 , and ⁇ V3 as ⁇ ave in the trajectory link candidate V.
- the ID association result totaling unit 34 may calculate the sum of T U1 and T U2 as T sum (total of detection time lengths) in the trajectory link candidate U in step S7014. Further, the ID association result totaling unit 34 may calculate the sum of T V1 , T V2 , and T V3 as T sum in the trajectory link candidate V.
- the ID association result totaling unit 34 normalizes ⁇ ave (average value of variance values of attribute information) obtained for each trajectory link candidate, and T sum (detection) obtained for each trajectory link candidate. Normalize the total time length).
- ID correspondence result counting section 34 a sigma ave individual trajectories linking candidates, respectively, may be divided by the sum of the sigma ave of each trajectory connecting candidate. Further, the ID association result totaling unit 34 may divide the T sum of individual trajectory link candidates by the sum of the T sums of the respective trajectory link candidates.
- the ID association result totaling unit 34 calculates the above-described equation (1) using ⁇ ave after normalization and T sum after normalization for each trajectory link candidate, and the result is calculated as the trajectory link.
- a score S representing the likelihood of the candidate is set (step S7015).
- Step S700 ends when step S7015 ends. Thereafter, the ID association result totaling unit 34 may perform the processing after step S701 shown in FIG.
- the ID association score can be calculated based on the score representing the likelihood of the trajectory link candidate obtained for each trajectory link candidate.
- trajectory connection A score representing the likelihood of the candidate may be calculated.
- the ID association result totaling unit 34 may calculate the following equation (2) for each trajectory link candidate, and use the result as a score S representing the likelihood of the trajectory link candidate.
- the normalized score is the score obtained in step S7008 of the first calculation method.
- the values of ⁇ , ⁇ , and ⁇ may be adjusted in advance so that good flow line information can be obtained.
- FIG. 26 is a block diagram showing an example of the minimum configuration of the flow line information generation system of the present invention.
- the flow line information generation system according to the present invention includes a trajectory link candidate generation unit 91, a likelihood calculation unit 92, an identification information association unit 93, an association tendency digitization unit 94, and an association score calculation unit 95. Prepare.
- the trajectory link candidate generation unit 91 (for example, the trajectory link candidate generation unit 31) is trajectory information representing a trajectory of the moving object or a fragment of the trajectory of the moving object due to the tracking interruption, and identifies the trajectory information.
- a trajectory of one moving object is obtained by using trajectory information including trajectory identification information (for example, trajectory number), position information of the moving object detected during tracking of the moving object, and a list of detection times of the position information.
- trajectory identification information for example, trajectory number
- a trajectory link that is a set of trajectory information that can be considered to be represented and a trajectory link candidate that is a set of trajectory links that include all trajectory information without omissions are generated.
- Likelihood calculation means 92 is a trajectory link / identification information pair that is a combination of trajectory link and mobile body identification information (mobile body ID) detected in the tracking area. For each (for example, trajectory link / ID pair), the likelihood (for example, ID association likelihood) representing the trajectory of the mobile object whose trajectory link is identified by the mobile object identification information is calculated.
- the identification information association unit 93 (for example, the ID association unit 33), for each trajectory link candidate, for each trajectory link candidate based on the likelihood calculated for each trajectory link / identification information pair. Corresponding mobile object identification information to
- the association tendency digitizing means 94 obtains the individual trajectory information from the result of the association between the trajectory links included in the individual trajectory link candidates and the mobile object identification information. For each individual mobile unit identification information, the tendency of the number of times that the trajectory information and the mobile unit identification information are associated is quantified.
- the association score calculation unit 95 (for example, the ID association score calculation unit 35) is a value indicating the tendency of the number of times that the trajectory information and the moving body identification information are associated (for example, each of the aggregation maps at the end of step S7). Based on the value of the cell), for each piece of trajectory information, a correspondence score indicating the degree of ambiguity of the correspondence relationship that the trajectory represented by the trajectory information corresponds to the trajectory of the mobile object identified by the mobile object identification information (for example, ID association score) is calculated for each mobile object identification information.
- association tendency digitizing means 94 refers to the result of association between the trajectory link and the moving object identification information obtained for each trajectory link candidate by the identification information associating means 93, and the trajectory link and the moving object identification information. 1 may be added to each set of individual trajectory information and moving body identification information in the trajectory link.
- the association tendency digitizing means 94 calculates a score representing the likelihood of the trajectory link candidate for each trajectory link candidate, and the trajectory link obtained for each trajectory link candidate by the identification information correlation means 93 and the moving object identification. With reference to the result of association with information, when a set in which the trajectory link and the mobile object identification information are associated is established, for each set of individual trajectory information and the mobile object identification information in the trajectory link, A score representing the likelihood of the trajectory link candidate including the trajectory link may be added.
- association tendency digitizing unit 94 calculates the likelihood regarding the combination of the trajectory link included in the trajectory link candidate and the mobile object identification information associated with the trajectory link by the identification information correlation unit 93.
- a score representing the likelihood of the trajectory link candidate may be calculated by multiplying for each trajectory link included in the candidate.
- trajectory link candidate generation unit 91 generates trajectory link and trajectory link candidates using the trajectory information including the attribute information of the moving object, and the association tendency digitizing unit 94 adds the trajectory link included in the trajectory link candidates.
- a score representing the likelihood of the trajectory link candidate may be calculated based on each attribute information of the mobile object included and the length from the first detection time to the last detection time in the trajectory information included in the trajectory link. .
- association score calculating means 95 may calculate the association score so that the sum of the association scores regarding each mobile object identification information becomes 1 for each trajectory information.
- trajectory link candidate generation means 91 may prohibit the trajectory link from a set of trajectory information corresponding to a predetermined condition.
- the present invention is preferably applied to a flow line information generation system for obtaining a correspondence relationship between a trajectory of a moving object and an ID.
- the flow line is detected by associating the position of the shopper who performs shopping in the shopping center, the member number unique to each shopper, and the ID association score, and the degree of likelihood of the shopper's flow line is visualized. Applicable to the system.
- the present invention can be applied to a system that visualizes a main flow line in which the flow lines of a plurality of persons with similar movement paths are collected for the purpose of marketing.
- the present invention is also applicable to a system for visualizing what kind of person is present at what ratio in the sales floor.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- General Business, Economics & Management (AREA)
- Radar, Positioning & Navigation (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
Description
ID“b”: 0.2+0.4=0.6
ID“c”: 0.2+0.4=0.6
ID“d”: 0.4
ID“e”: 0.4
図7は、動線情報生成部3の処理経過の例を示すフローチャートである。軌跡連結候補生成手段31は、軌跡情報入力部1から、軌跡番号、移動体の位置座標およびその検出時刻の組を取得する(ステップS1)。軌跡連結候補生成手段31は、逐次入力される軌跡番号、位置座標および検出時刻の組を、軌跡番号毎に、過去一定時間分の検出時刻と位置座標のリストを持つようなデータ構造として記憶する。例えば、時刻t1からtnまでの軌跡番号1に関する情報は、(軌跡番号1,(時刻t1,x座標,y座標),(時刻t2,x座標,y座標),・・・,(時刻tn,x座標,y座標))等のように表される。すなわち、軌跡連結候補生成手段31は、軌跡番号と、過去一定時間分の検出時刻および検出座標のリストとを含む軌跡情報を生成し、記憶する。具体的には、新たに、軌跡番号、位置座標および検出時刻の組が入力されたときに、軌跡連結候補生成手段31は、その軌跡番号に対応するリストに、新たに入力された位置座標および検出時刻の組を追加すればよい。また、軌跡連結候補生成手段31は、過去一定時間よりも前の検出時刻および位置座標については、リストから削除すればよい。
α・(正規化後のσave)+β・(正規化後のTsum)+γ・(正規化後のスコア) 式(2)
2 ID情報入力部
3 動線情報生成部
4 動線情報分析部
31 軌跡連結候補生成手段
32 ID対応付け尤度算出手段
33 ID対応付け手段
34 ID対応付け結果集計手段
35 ID対応付けスコア算出手段
36 スコア付きID候補出力手段
Claims (13)
- 移動体の軌跡または追跡の途切れが生じたことによる移動体の軌跡の断片を表す軌跡情報であって、軌跡情報を識別する軌跡識別情報と、移動体の追跡中に検出された移動体の位置情報および当該位置情報の検出時刻のリストとを含む軌跡情報を用いて、1つの移動体の軌跡を表しているとみなすことができる軌跡情報の集合である軌跡連結と、全ての軌跡情報を漏れなく重複なく含んだ軌跡連結の集合である軌跡連結候補とを生成する軌跡連結候補生成手段と、
軌跡連結と追跡領域内で検出された移動体識別情報との組み合わせである軌跡連結・識別情報ペア毎に、前記軌跡連結が前記移動体識別情報によって識別される移動体の軌跡を表している尤度を算出する尤度算出手段と、
軌跡連結・識別情報ペア毎に算出された前記尤度に基づいて、軌跡連結候補毎に、軌跡連結候補に含まれる個々の軌跡連結に尤もらしい移動体識別情報を対応付ける識別情報対応付け手段と、
個々の軌跡連結候補に含まれる各軌跡連結と移動体識別情報との対応付けの結果から、個々の軌跡情報に関して、個々の移動体識別情報毎に、軌跡情報と移動体識別情報とが対応付けられた回数の傾向を数値化する対応付け傾向数値化手段と、
軌跡情報と移動体識別情報とが対応付けられた回数の傾向を示す値に基づいて、軌跡情報毎に、軌跡情報が表す軌跡が移動体識別情報によって識別される移動体の軌跡に該当するという対応関係のあいまいさの程度を示す対応付けスコアを各移動体識別情報に関して算出する対応付けスコア算出手段とを備える
ことを特徴とする動線情報生成システム。 - 対応付け傾向数値化手段は、識別情報対応付け手段によって軌跡連結候補毎に得られた軌跡連結と移動体識別情報との対応付け結果を参照し、軌跡連結と移動体識別情報とを対応付けた組が成立している場合に、前記軌跡連結内の個々の軌跡情報と前記移動体識別情報との各組に関して1を加算する
請求項1に記載の動線情報生成システム。 - 対応付け傾向数値化手段は、
軌跡連結候補毎に軌跡連結候補の尤もらしさを表すスコアを算出し、
識別情報対応付け手段によって軌跡連結候補毎に得られた軌跡連結と移動体識別情報との対応付け結果を参照し、軌跡連結と移動体識別情報とを対応付けた組が成立している場合に、前記軌跡連結内の個々の軌跡情報と前記移動体識別情報との各組に関して、前記軌跡連結を含む軌跡連結候補の尤もらしさを表すスコアを加算する
請求項1に記載の動線情報生成システム。 - 対応付け傾向数値化手段は、
軌跡連結候補に含まれる軌跡連結と、識別情報対応付け手段によって当該軌跡連結に対応付けられた移動体識別情報との組に関する尤度を、前記軌跡連結候補に含まれる軌跡連結毎に乗算することによって、軌跡連結候補の尤もらしさを表すスコアを算出する
請求項3に記載の動線情報生成システム。 - 軌跡連結候補生成手段は、移動体の属性情報を含む軌跡情報を用いて軌跡連結および軌跡連結候補を生成し、
対応付け傾向数値化手段は、
軌跡連結候補に含まれる軌跡連結に含まれる移動体の各属性情報と、軌跡連結に含まれる軌跡情報における最初の検出時刻から最後の検出時刻までの長さとに基づいて、軌跡連結候補の尤もらしさを表すスコアを算出する
請求項3に記載の動線情報生成システム。 - 対応付けスコア算出手段は、軌跡情報毎に各移動体識別情報に関する対応付けスコアの総和が1になるように対応付けスコアを算出する
請求項1から請求項5のうちのいずれか1項に記載の動線情報生成システム。 - 軌跡連結候補生成手段は、予め定められた所定の条件に該当する軌跡情報の集合を軌跡連結とすることを禁止する
請求項1から請求項6のうちのいずれか1項に記載の動線情報生成システム。 - 移動体の軌跡または追跡の途切れが生じたことによる移動体の軌跡の断片を表す軌跡情報であって、軌跡情報を識別する軌跡識別情報と、移動体の追跡中に検出された移動体の位置情報および当該位置情報の検出時刻のリストとを含む軌跡情報を用いて、1つの移動体の軌跡を表しているとみなすことができる軌跡情報の集合である軌跡連結と、全ての軌跡情報を漏れなく重複なく含んだ軌跡連結の集合である軌跡連結候補とを生成し、
軌跡連結と追跡領域内で検出された移動体識別情報との組み合わせである軌跡連結・識別情報ペア毎に、前記軌跡連結が前記移動体識別情報によって識別される移動体の軌跡を表している尤度を算出し、
軌跡連結・識別情報ペア毎に算出された前記尤度に基づいて、軌跡連結候補毎に、軌跡連結候補に含まれる個々の軌跡連結に尤もらしい移動体識別情報を対応付け、
個々の軌跡連結候補に含まれる各軌跡連結と移動体識別情報との対応付けの結果から、個々の軌跡情報に関して、個々の移動体識別情報毎に、軌跡情報と移動体識別情報とが対応付けられた回数の傾向を数値化し、
軌跡情報と移動体識別情報とが対応付けられた回数の傾向を示す値に基づいて、軌跡情報毎に、軌跡情報が表す軌跡が移動体識別情報によって識別される移動体の軌跡に該当するという対応関係のあいまいさの程度を示す対応付けスコアを各移動体識別情報に関して算出する
ことを特徴とする動線情報生成方法。 - 軌跡連結候補毎に得られた軌跡連結と移動体識別情報との対応付け結果を参照し、軌跡連結と移動体識別情報とを対応付けた組が成立している場合に、前記軌跡連結内の個々の軌跡情報と前記移動体識別情報との各組に関して1を加算する
請求項8に記載の動線情報生成方法。 - 軌跡連結候補毎に軌跡連結候補の尤もらしさを表すスコアを算出し、
軌跡連結候補毎に得られた軌跡連結と移動体識別情報との対応付け結果を参照し、軌跡連結と移動体識別情報とを対応付けた組が成立している場合に、前記軌跡連結内の個々の軌跡情報と前記移動体識別情報との各組に関して、前記軌跡連結を含む軌跡連結候補の尤もらしさを表すスコアを加算する
請求項8に記載の動線情報生成方法。 - コンピュータに、
移動体の軌跡または追跡の途切れが生じたことによる移動体の軌跡の断片を表す軌跡情報であって、軌跡情報を識別する軌跡識別情報と、移動体の追跡中に検出された移動体の位置情報および当該位置情報の検出時刻のリストとを含む軌跡情報を用いて、1つの移動体の軌跡を表しているとみなすことができる軌跡情報の集合である軌跡連結と、全ての軌跡情報を漏れなく重複なく含んだ軌跡連結の集合である軌跡連結候補とを生成する軌跡連結候補生成処理、
軌跡連結と追跡領域内で検出された移動体識別情報との組み合わせである軌跡連結・識別情報ペア毎に、前記軌跡連結が前記移動体識別情報によって識別される移動体の軌跡を表している尤度を算出する尤度算出処理、
軌跡連結・識別情報ペア毎に算出された前記尤度に基づいて、軌跡連結候補毎に、軌跡連結候補に含まれる個々の軌跡連結に尤もらしい移動体識別情報を対応付ける識別情報対応付け処理、
個々の軌跡連結候補に含まれる各軌跡連結と移動体識別情報との対応付けの結果から、個々の軌跡情報に関して、個々の移動体識別情報毎に、軌跡情報と移動体識別情報とが対応付けられた回数の傾向を数値化する対応付け傾向数値化処理、および、
軌跡情報と移動体識別情報とが対応付けられた回数の傾向を示す値に基づいて、軌跡情報毎に、軌跡情報が表す軌跡が移動体識別情報によって識別される移動体の軌跡に該当するという対応関係のあいまいさの程度を示す対応付けスコアを各移動体識別情報に関して算出する対応付けスコア算出処理
を実行させるための動線情報生成プログラム。 - コンピュータに、
対応付け傾向数値化処理で、識別情報対応付け処理によって軌跡連結候補毎に得られた軌跡連結と移動体識別情報との対応付け結果を参照し、軌跡連結と移動体識別情報とを対応付けた組が成立している場合に、前記軌跡連結内の個々の軌跡情報と前記移動体識別情報との各組に関して1を加算させる
請求項11に記載の動線情報生成プログラム。 - 対応付け傾向数値化処理で
軌跡連結候補毎に軌跡連結候補の尤もらしさを表すスコアを算出させ、
識別情報対応付け処理によって軌跡連結候補毎に得られた軌跡連結と移動体識別情報との対応付け結果を参照し、軌跡連結と移動体識別情報とを対応付けた組が成立している場合に、前記軌跡連結内の個々の軌跡情報と前記移動体識別情報との各組に関して、前記軌跡連結を含む軌跡連結候補の尤もらしさを表すスコアを加算させる
請求項11に記載の動線情報生成プログラム。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US14/380,077 US10648803B2 (en) | 2012-02-29 | 2013-02-21 | Movement line information generation system, movement line information generation method and movement line information generation program |
JP2014502011A JP6070689B2 (ja) | 2012-02-29 | 2013-02-21 | 動線情報生成システム、動線情報生成方法および動線情報生成プログラム |
US16/509,715 US10895454B2 (en) | 2012-02-29 | 2019-07-12 | Movement line information generation system, movement line information generation method and movement line information generation program |
US16/509,732 US20190331485A1 (en) | 2012-02-29 | 2019-07-12 | Movement line information generation system, movement line information generation method and movement line information generation program |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2012-043641 | 2012-02-29 | ||
JP2012043641 | 2012-02-29 |
Related Child Applications (3)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/380,077 A-371-Of-International US10648803B2 (en) | 2012-02-29 | 2013-02-21 | Movement line information generation system, movement line information generation method and movement line information generation program |
US16/509,715 Continuation US10895454B2 (en) | 2012-02-29 | 2019-07-12 | Movement line information generation system, movement line information generation method and movement line information generation program |
US16/509,732 Continuation US20190331485A1 (en) | 2012-02-29 | 2019-07-12 | Movement line information generation system, movement line information generation method and movement line information generation program |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2013128852A1 true WO2013128852A1 (ja) | 2013-09-06 |
Family
ID=49082067
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2013/000976 WO2013128852A1 (ja) | 2012-02-29 | 2013-02-21 | 動線情報生成システム、動線情報生成方法および動線情報生成プログラム |
Country Status (3)
Country | Link |
---|---|
US (3) | US10648803B2 (ja) |
JP (1) | JP6070689B2 (ja) |
WO (1) | WO2013128852A1 (ja) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2015070354A (ja) * | 2013-09-27 | 2015-04-13 | パナソニックIpマネジメント株式会社 | 移動体追跡装置、移動体追跡システムおよび移動体追跡方法 |
JP2015201005A (ja) * | 2014-04-07 | 2015-11-12 | パナソニック株式会社 | 軌跡解析装置および軌跡解析方法 |
JP2016085487A (ja) * | 2014-10-22 | 2016-05-19 | キヤノン株式会社 | 情報処理装置、情報処理方法及びコンピュータプログラム |
WO2016084498A1 (ja) * | 2014-11-28 | 2016-06-02 | 古野電気株式会社 | 追尾処理装置及び追尾処理方法 |
JP2016129309A (ja) * | 2015-01-09 | 2016-07-14 | 富士通株式会社 | オブジェクト紐付方法、装置、及びプログラム |
JP2017083980A (ja) * | 2015-10-26 | 2017-05-18 | 日立アイ・エヌ・エス・ソフトウェア株式会社 | 行動分析装置及びシステム及び方法 |
WO2018051507A1 (ja) * | 2016-09-16 | 2018-03-22 | 株式会社日立情報通信エンジニアリング | 物体識別システム及び物体識別方法 |
JP2018048816A (ja) * | 2016-09-20 | 2018-03-29 | 株式会社東芝 | 学習データ作成装置、学習データ作成方法及びコンピュータプログラム |
WO2018092881A1 (ja) * | 2016-11-21 | 2018-05-24 | 日本電気株式会社 | 動線合成装置、動線合成方法および記録媒体 |
WO2018163804A1 (ja) * | 2017-03-06 | 2018-09-13 | キヤノン株式会社 | 情報処理システム、情報処理装置、情報処理方法ならびに情報処理方法をコンピュータに実行させるプログラム |
WO2019138537A1 (ja) * | 2018-01-12 | 2019-07-18 | 日本電気株式会社 | 情報処理装置、制御方法、及びプログラム |
JP2020091174A (ja) * | 2018-12-05 | 2020-06-11 | パナソニックIpマネジメント株式会社 | 推定識別装置および推定識別システム |
JPWO2021065927A1 (ja) * | 2019-10-02 | 2021-04-08 | ||
JP2021071296A (ja) * | 2019-10-29 | 2021-05-06 | 三菱電機ビルテクノサービス株式会社 | 無線タグ位置検出システム |
US11189038B2 (en) | 2017-11-21 | 2021-11-30 | Mitsubishi Electric Corporation | Tracking apparatus and computer readable medium |
WO2022254890A1 (ja) * | 2021-06-04 | 2022-12-08 | パナソニックIpマネジメント株式会社 | 動線解析装置、動線解析方法及びコンピュータプログラム |
WO2023132090A1 (ja) * | 2022-01-07 | 2023-07-13 | パナソニックIpマネジメント株式会社 | データ分析装置および方法 |
WO2024101472A1 (ko) * | 2022-11-09 | 2024-05-16 | 주식회사 휴먼아이씨티 | 영상 내에서 객체를 처리하는 방법 및 장치 |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB201512283D0 (en) * | 2015-07-14 | 2015-08-19 | Apical Ltd | Track behaviour events |
US10127417B2 (en) | 2016-01-08 | 2018-11-13 | Disney Enterprises, Inc. | Systems and methods for determining interaction states of wireless tags based on probability distributions |
US11048927B2 (en) | 2017-10-24 | 2021-06-29 | Waymo Llc | Pedestrian behavior predictions for autonomous vehicles |
JP7102790B2 (ja) * | 2018-03-07 | 2022-07-20 | 富士通株式会社 | データ生成方法、データ生成プログラム及び情報処理装置 |
CA3106694A1 (en) * | 2018-07-18 | 2020-01-23 | Robert Laganiere | System and method for tracking customer movements in a customer service environment |
JP2020102135A (ja) * | 2018-12-25 | 2020-07-02 | 清水建設株式会社 | 追跡システム、追跡処理方法、及びプログラム |
CN114791783B (zh) * | 2022-03-18 | 2024-09-17 | 联想(北京)有限公司 | 一种信息处理方法和信息处理设备 |
CN116363565B (zh) * | 2023-06-01 | 2023-08-11 | 深圳须弥云图空间科技有限公司 | 目标轨迹确定方法、装置、电子设备及存储介质 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005031955A (ja) * | 2003-07-11 | 2005-02-03 | Kddi Corp | 移動体追跡システム |
JP2007240295A (ja) * | 2006-03-08 | 2007-09-20 | Advanced Telecommunication Research Institute International | 位置推定装置、位置推定方法及び位置推定プログラム |
JP2008122093A (ja) * | 2006-11-08 | 2008-05-29 | Mitsubishi Electric Corp | 多目標追尾装置 |
WO2010095437A1 (ja) * | 2009-02-19 | 2010-08-26 | パナソニック株式会社 | 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 |
WO2011021588A1 (ja) * | 2009-08-20 | 2011-02-24 | 日本電気株式会社 | 移動体軌跡識別システム |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
IL114838A0 (en) | 1995-08-04 | 1996-11-14 | Spiegel Ehud | Apparatus and method for object tracking |
JP4697500B2 (ja) | 1999-08-09 | 2011-06-08 | ソニー株式会社 | 送信装置および送信方法、受信装置および受信方法、並びに記録媒体 |
DE60134210D1 (de) | 2000-12-27 | 2008-07-10 | Fujifilm Corp | System und Verfahren zur Informationsbenachrichtigung |
JP4576060B2 (ja) * | 2001-02-15 | 2010-11-04 | 富士通株式会社 | 自動契約装置および自動契約方法ならびに自動契約プログラム |
US8409024B2 (en) * | 2001-09-12 | 2013-04-02 | Pillar Vision, Inc. | Trajectory detection and feedback system for golf |
US8149278B2 (en) * | 2006-11-30 | 2012-04-03 | Mitsubishi Electric Research Laboratories, Inc. | System and method for modeling movement of objects using probabilistic graphs obtained from surveillance data |
US20090195401A1 (en) * | 2008-01-31 | 2009-08-06 | Andrew Maroney | Apparatus and method for surveillance system using sensor arrays |
US8253564B2 (en) * | 2009-02-19 | 2012-08-28 | Panasonic Corporation | Predicting a future location of a moving object observed by a surveillance device |
CN102939609B (zh) * | 2009-12-16 | 2016-08-03 | 汤姆森特许公司 | 基于人类交互轨迹的系统 |
US8825388B2 (en) | 2010-07-13 | 2014-09-02 | Qualcomm Incorporated | Indoor likelihood heatmap |
-
2013
- 2013-02-21 JP JP2014502011A patent/JP6070689B2/ja active Active
- 2013-02-21 US US14/380,077 patent/US10648803B2/en active Active
- 2013-02-21 WO PCT/JP2013/000976 patent/WO2013128852A1/ja active Application Filing
-
2019
- 2019-07-12 US US16/509,732 patent/US20190331485A1/en not_active Abandoned
- 2019-07-12 US US16/509,715 patent/US10895454B2/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005031955A (ja) * | 2003-07-11 | 2005-02-03 | Kddi Corp | 移動体追跡システム |
JP2007240295A (ja) * | 2006-03-08 | 2007-09-20 | Advanced Telecommunication Research Institute International | 位置推定装置、位置推定方法及び位置推定プログラム |
JP2008122093A (ja) * | 2006-11-08 | 2008-05-29 | Mitsubishi Electric Corp | 多目標追尾装置 |
WO2010095437A1 (ja) * | 2009-02-19 | 2010-08-26 | パナソニック株式会社 | 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 |
WO2011021588A1 (ja) * | 2009-08-20 | 2011-02-24 | 日本電気株式会社 | 移動体軌跡識別システム |
Non-Patent Citations (1)
Title |
---|
YUKIE MORIGUCHI: "A Robust and Fast Human Trajectory Detection Method based on Sensor Fusion", IEICE TECHNICAL REPORT, vol. 110, no. 454, 28 February 2011 (2011-02-28), pages 7 - 14 * |
Cited By (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2015070354A (ja) * | 2013-09-27 | 2015-04-13 | パナソニックIpマネジメント株式会社 | 移動体追跡装置、移動体追跡システムおよび移動体追跡方法 |
JP2015201005A (ja) * | 2014-04-07 | 2015-11-12 | パナソニック株式会社 | 軌跡解析装置および軌跡解析方法 |
JP2016085487A (ja) * | 2014-10-22 | 2016-05-19 | キヤノン株式会社 | 情報処理装置、情報処理方法及びコンピュータプログラム |
US10585184B2 (en) | 2014-11-28 | 2020-03-10 | Furuno Electric Co., Ltd. | Tracking processor and method of tracking processing |
WO2016084498A1 (ja) * | 2014-11-28 | 2016-06-02 | 古野電気株式会社 | 追尾処理装置及び追尾処理方法 |
CN107110967A (zh) * | 2014-11-28 | 2017-08-29 | 古野电气株式会社 | 追踪处理装置及追踪处理方法 |
JPWO2016084498A1 (ja) * | 2014-11-28 | 2017-09-07 | 古野電気株式会社 | 追尾処理装置及び追尾処理方法 |
CN107110967B (zh) * | 2014-11-28 | 2020-10-16 | 古野电气株式会社 | 追踪处理装置及追踪处理方法 |
JP2016129309A (ja) * | 2015-01-09 | 2016-07-14 | 富士通株式会社 | オブジェクト紐付方法、装置、及びプログラム |
JP2017083980A (ja) * | 2015-10-26 | 2017-05-18 | 日立アイ・エヌ・エス・ソフトウェア株式会社 | 行動分析装置及びシステム及び方法 |
WO2018051507A1 (ja) * | 2016-09-16 | 2018-03-22 | 株式会社日立情報通信エンジニアリング | 物体識別システム及び物体識別方法 |
JP2018048816A (ja) * | 2016-09-20 | 2018-03-29 | 株式会社東芝 | 学習データ作成装置、学習データ作成方法及びコンピュータプログラム |
WO2018092881A1 (ja) * | 2016-11-21 | 2018-05-24 | 日本電気株式会社 | 動線合成装置、動線合成方法および記録媒体 |
US11023739B2 (en) | 2016-11-21 | 2021-06-01 | Nec Corporation | Flow line combining device, flow line combining method, and recording medium |
JPWO2018092881A1 (ja) * | 2016-11-21 | 2019-08-08 | 日本電気株式会社 | 動線合成装置、動線合成方法およびプログラム |
US11157747B2 (en) | 2017-03-06 | 2021-10-26 | Canon Kabushiki Kaisha | Information-processing system, information-processing apparatus, method of processing information, and storage medium storing program for causing computer to execute method of processing information |
WO2018163804A1 (ja) * | 2017-03-06 | 2018-09-13 | キヤノン株式会社 | 情報処理システム、情報処理装置、情報処理方法ならびに情報処理方法をコンピュータに実行させるプログラム |
US11189038B2 (en) | 2017-11-21 | 2021-11-30 | Mitsubishi Electric Corporation | Tracking apparatus and computer readable medium |
JPWO2019138537A1 (ja) * | 2018-01-12 | 2020-12-17 | 日本電気株式会社 | 情報処理装置、制御方法、及びプログラム |
WO2019138537A1 (ja) * | 2018-01-12 | 2019-07-18 | 日本電気株式会社 | 情報処理装置、制御方法、及びプログラム |
US11803972B2 (en) | 2018-01-12 | 2023-10-31 | Nec Corporation | Information processing apparatus, control method, and program for accurately linking fragmented trajectories of an object |
JP2020091174A (ja) * | 2018-12-05 | 2020-06-11 | パナソニックIpマネジメント株式会社 | 推定識別装置および推定識別システム |
JP7266237B2 (ja) | 2018-12-05 | 2023-04-28 | パナソニックIpマネジメント株式会社 | 推定識別装置および推定識別システム |
US11940548B2 (en) | 2018-12-05 | 2024-03-26 | Panasonic Intellectual Property Management Co., Ltd. | Estimation and identification device and estimation and identification system |
WO2021065927A1 (ja) * | 2019-10-02 | 2021-04-08 | 株式会社Nttドコモ | データ補正装置 |
JPWO2021065927A1 (ja) * | 2019-10-02 | 2021-04-08 | ||
JP2021071296A (ja) * | 2019-10-29 | 2021-05-06 | 三菱電機ビルテクノサービス株式会社 | 無線タグ位置検出システム |
JP7325297B2 (ja) | 2019-10-29 | 2023-08-14 | 三菱電機ビルソリューションズ株式会社 | 無線タグ位置検出システム |
WO2022254890A1 (ja) * | 2021-06-04 | 2022-12-08 | パナソニックIpマネジメント株式会社 | 動線解析装置、動線解析方法及びコンピュータプログラム |
WO2023132090A1 (ja) * | 2022-01-07 | 2023-07-13 | パナソニックIpマネジメント株式会社 | データ分析装置および方法 |
WO2024101472A1 (ko) * | 2022-11-09 | 2024-05-16 | 주식회사 휴먼아이씨티 | 영상 내에서 객체를 처리하는 방법 및 장치 |
Also Published As
Publication number | Publication date |
---|---|
US20190331484A1 (en) | 2019-10-31 |
US10648803B2 (en) | 2020-05-12 |
US20190331485A1 (en) | 2019-10-31 |
JPWO2013128852A1 (ja) | 2015-07-30 |
US20150019161A1 (en) | 2015-01-15 |
JP6070689B2 (ja) | 2017-02-01 |
US10895454B2 (en) | 2021-01-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6070689B2 (ja) | 動線情報生成システム、動線情報生成方法および動線情報生成プログラム | |
JP5682563B2 (ja) | 移動体軌跡識別システム、移動体軌跡識別方法および移動体軌跡識別プログラム | |
Sattarian et al. | Indoor navigation systems based on data mining techniques in internet of things: a survey | |
US10387896B1 (en) | At-shelf brand strength tracking and decision analytics | |
Taneja et al. | Analysis of three indoor localization technologies for supporting operations and maintenance field tasks | |
KR101369261B1 (ko) | 스마트 단말을 이용한 사용자의 이동 경로 실시간 예측 방법 및 그 시스템 | |
Trogh et al. | Advanced real-time indoor tracking based on the viterbi algorithm and semantic data | |
US20170146349A1 (en) | Landmark location determination | |
CN101526601A (zh) | 采取toa与rss融合方式的自适应定位方法、设备和系统 | |
JP2015531053A (ja) | 無線マップを動的に作成するためのシステム、方法、及びコンピュータプログラム | |
US8731829B2 (en) | Flow line detection system, flow line detection method, and flow line detection program | |
US9515752B2 (en) | Method for mapping wi-fi fingerprints by means of blind collection of RSS measurements | |
Guo et al. | Shopprofiler: Profiling shops with crowdsourcing data | |
Scholz et al. | Resource-aware on-line RFID localization using proximity data | |
Zhang et al. | Unambiguous association of crowd-sourced radio maps to floor plans for indoor localization | |
JP2013217802A (ja) | 軌跡解析装置および軌跡解析プログラム | |
GB2570853A (en) | Identifying sites visited by a user device | |
CN113108806A (zh) | 路径规划方法、装置、设备及介质 | |
Xu et al. | Dccla: Automatic indoor localization using unsupervised wi-fi fingerprinting | |
Wietrzykowski et al. | Adopting the FAB-MAP algorithm for indoor localization with WiFi fingerprints | |
Wang et al. | Received signal strength–based localization for large space indoor environments | |
Kim et al. | Cooperative localization considering estimated location uncertainty in distributed ad hoc networks | |
WO2023220091A1 (en) | Monitored environment fingerprint modeling system | |
Haznedar et al. | Hybrid indoor positioning for smart homes using WiFi and Bluetooth low energy technologies | |
Kerdjidj et al. | Exploring 2D Representation and Transfer Learning Techniques for People Identification in Indoor Localization |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 13755803 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2014502011 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14380077 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 13755803 Country of ref document: EP Kind code of ref document: A1 |