Mobile body tracking system

Abstract

To provide a configuration capable of tracking a moving path of a mobile body without acquiring feature information of a detected object.SOLUTION: Tracking position information of a mobile body is registered in a tracking information list in association with identification information and detection timing. Further, a moving destination to which the mobile body moves next is predicted as a movement prediction position for each identification information by using the tracking information list, and movement prediction position information is associated with the identification information and registered in the tracking information list. Then, an object detection position is included in a movement prediction range E set according to the movement prediction position information, so that the mobile body is specified, and the information related to the object detection position is registered in the tracking information list as tracking position information associated with the same identification information as the movement prediction position information.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a moving object tracking system that tracks and acquires the movement path of a moving object that has entered a predetermined monitoring area. [Background technology]

[0002] Conventionally, as a mobile object tracking system that tracks and acquires the movement path of a mobile object that has entered a specified monitoring area, there is known an entrance / exit management system as disclosed in the following Patent Document 1. In this entrance / exit management system, an IC card reader and a camera are placed at the entrance / exit to be managed, and based on the reading result by the IC card reader and image data from the camera, it is determined whether or not the movement path of the user who has had the IC card read differs from the path that passes through the entrance / exit, thereby controlling the entrance / exit management of users. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2017-037447 A Summary of the Invention [Problem to be solved by the invention]

[0004] As described above, in order to track a moving object, it is necessary to know that the object detected this time and the object detected last time are the same individual. In a system that uses a camera to capture an image of a monitoring area as described above, the color and dimensions of the detected object are acquired as feature information, and objects with similar colors and dimensions are regarded as the same individual, making it possible to determine that the moving object has moved from the position of the object detected last time to the position of the object detected this time.

[0005] In a moving object tracking system that uses LiDAR or other devices to detect the position of an object by receiving light reflected from the object, the moving object's path can be tracked by detecting the object's previous and next positions.

[0006] However, with LiDAR and the like, it is difficult to obtain characteristic information such as the color and dimensions of a detected object, and therefore it may not be possible to accurately determine that the currently detected object and the previously detected object are the same individual. For this reason, for example, when a moving object is tracked so that the closest detected detection position is set as the next moving destination, there are cases in which the moving object cannot be tracked accurately because two moving objects temporarily come close to each other, resulting in a mismatch in the detection position to be tracked.

[0007] The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a configuration that can track the movement path of a moving object without acquiring characteristic information of the detected object. [Means for solving the problem]

[0008] In order to achieve the above object, the present invention provides A moving object tracking system (1) for tracking a moving object that has entered a predetermined monitoring area (S), a detection unit (10) that detects a position of an object as an object detection position (Px) by receiving reflected light from an object, including a moving object, within the specified monitoring area in response to light projection toward the specified monitoring area; a storage unit (22) having a tracking information list in which tracking position information regarding a tracking position (Po) of a moving object is registered in association with identification information for distinguishing the moving object from other moving objects and detection timing; a prediction unit (21) that predicts a next destination of the moving object as a movement predicted position (Pp) for each of the identification information by using the tracking information list, and registers movement predicted position information related to the movement predicted position in the tracking information list in association with the identification information; an identification unit (21) that identifies a moving object when the object detection position is included within a movement prediction range (E) that is set according to the movement prediction position information, and registers information regarding the object detection position in the tracking information list as the tracking position information associated with the same identification information as the movement prediction position information; a tracking unit (21) that tracks the movement of a moving object within the specified monitoring area based on the information registered in the tracking information list; The present invention is characterized by comprising: The symbols in parentheses above indicate the corresponding relationship with the specific means described in the embodiments described later. Effect of the Invention

[0009] In the present invention, tracking position information of a moving object is associated with identification information and detection timing and registered in a tracking information list. The prediction unit uses the tracking information list to predict the next destination of the moving object for each identification information as a movement predicted position, and registers the movement predicted position information in association with the identification information in the tracking information list. The identification unit identifies the moving object when the object detection position is included within a movement prediction range set according to the movement predicted position information, and registers information regarding the object detection position in the tracking information list as tracking position information associated with the same identification information as the movement predicted position information.

[0010] This makes it possible to identify the next object detection position based on the information registered in the tracking information list, i.e., the predicted movement range predicted from the previous tracking positions of the moving object, thereby making it possible to track the movement path of the moving object without obtaining characteristic information such as the color or dimensions of the detected object.

[0011] When multiple object detection positions are included within one movement prediction range, the identification unit may calculate candidate vectors from the tracking position that is the prediction source of the movement prediction range to the object detection position, and also calculate a predicted vector from the same tracking position to the movement prediction position, and register information regarding the object detection position of the candidate vector that is most similar to the predicted vector among the multiple candidate vectors in the tracking information list as tracking position information.

[0012] As a result, even if multiple object detection positions are included within the predicted movement range due to multiple moving bodies temporarily coming into close proximity, the next object detection position to which the object will move can be identified by taking into account the direction of movement, etc., thereby improving the tracking accuracy of moving bodies.

[0013] An object detection position that is not included within the predicted movement range is registered in the tentative tracking information list as a tentative tracking position, and when a determination is made a predetermined number of times for the same object that the object detection position is away from the previously registered tentative tracking position by a predetermined distance, the identification information assignment unit may associate information regarding the object detection position with newly assigned identification information and register it in the tracking information list as tracking position information.

[0014] This prevents identification information from being unnecessarily assigned to object detection positions that are not located a predetermined distance away from the previously registered tentative tracking position, object detection positions that are not continuously detected, and object detection positions that are temporarily detected, such as noise, etc. This reduces the processing load for tracking the moving object.

[0015] The flag setting unit may set a moving flag for identification information determined to be moving and a stopped flag for identification information determined to be stopped based on the relationship between the object detection position and the predicted movement range, and the deletion unit may delete identification information for which the stopped flag has been set for a certain continuous period of time from the tracking information list together with associated information.

[0016] As a result, if a moving object moves into a blind spot of the detection unit and becomes untraceable, the stopped flag is set for a certain period of time, and the identification information of the moving object is deleted from the tracking information list. Therefore, there is no unnecessary movement prediction for the identification information of an untraceable moving object, and the processing load for tracking the moving object can be reduced. In addition, if a moving object that has been temporarily stopped moves again, the flag of the moving object is reset to a moving flag, and tracking processing can be continued for the moving object.

[0017] The vehicle may be provided with a total number calculation unit that calculates the total number of moving objects that have entered a specified monitoring area using the number of identification information newly assigned by the identification information assignment unit and the number of identification information for which a stopped flag has been set by the flag setting unit.

[0018] For example, if three moving objects are being tracked and the total number of moving objects that have entered a specified monitoring area is calculated as "3," when one moving object moves again after temporarily stopping, the total number may be erroneously calculated as "4" as a new moving object is detected. For this reason, the number of identification information items with a stop flag set (hereinafter, referred to as the number of stopped moving object deletions) is recorded, and the number of stopped moving object deletions is reduced when new identification information is assigned. As a result, in the above example, even if one moving object moves again, the number of stopped moving object deletions is counted because the moving object was temporarily stopped, so by taking this number of stopped moving object deletions into consideration, the total number can be correctly calculated as "3." [Brief description of the drawings]

[0019] [Figure 1] FIG. 1 is an explanatory diagram illustrating a monitoring area in which a moving object is tracked by a moving object tracking system according to a first embodiment. [Diagram 2] 1 is a block diagram showing a schematic configuration of a moving object tracking system according to a first embodiment. [Diagram 3] FIG. 3 is an explanatory diagram illustrating the positional relationship between a predicted movement range predicted using information registered in a tracking information list and an object detection position that is a candidate movement destination, where FIG. 3(A) shows a case where one object detection position is included in the movement prediction range, and FIG. 3(B) shows a case where two object detection positions are included in the movement prediction range. [Figure 4] 10 is an explanatory diagram illustrating a candidate vector and a predicted vector calculated when the same two object detection positions are included in two predicted movement ranges, respectively. FIG. [Diagram 5] 10 is an explanatory diagram illustrating an example of a similarity calculated when the same two object detection positions are included in two predicted movement ranges. FIG. [Figure 6] 4 is a part of a flowchart illustrating the flow of a moving object tracking process performed by a control unit in the first embodiment. [Figure 7] 4 is a part of a flowchart illustrating the flow of a moving object tracking process performed by a control unit in the first embodiment. [Figure 8] 8A and 8B are explanatory diagrams illustrating the main parts of a moving object tracking system according to a second embodiment, in which FIG. 8A shows a case in which identification information is assigned, and FIG. 8B and FIG. 8C show cases in which identification information is not assigned. [Figure 9] 13 is a part of a flowchart illustrating the flow of a moving object tracking process performed by a control unit in the second embodiment. [Figure 10] 13 is a part of a flowchart illustrating the flow of a moving object tracking process performed by a control unit in the third embodiment. [Figure 11] 13 is a part of a flowchart illustrating the flow of a moving object tracking process performed by a control unit in the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] [First embodiment] Hereinafter, a first embodiment of a moving object tracking system according to the present invention will be described with reference to the drawings. The moving object tracking system 1 according to this embodiment is a system that tracks the movement of moving objects such as people who enter a predetermined monitoring area (hereinafter also referred to as monitoring area S). In this embodiment, as illustrated in Fig. 1, a room with three entrances Sa, Sb, and Sc is set as the monitoring area S, and a person who enters the room through any of the entrances is tracked, and tracking continues until the person leaves the room through any of the entrances.

[0021] As shown in FIG. 2, the moving object tracking system 1 includes a detection unit 10 that detects the position of an object within a monitoring area S, and a tracking device 20 that tracks the movement of a person (moving object) within the monitoring area S based on the detection result by the detection unit 10.

[0022] The detection unit 10 is a two-dimensional scanning distance measuring device such as LiDAR (Light Detection and Ranging), and is configured to measure the distance to an object (a position based on the detection unit 10) by receiving reflected light from the object in the irradiation direction of the measurement light. The detection unit 10 is disposed at a predetermined position within a monitoring area S as shown in FIG. 1, and functions to output the detection result to the tracking device 20.

[0023] As shown in FIG. 2, the tracking device 20 includes a control unit 21 for overall control and a memory unit 22 consisting of a semiconductor memory or the like, as well as a display unit 23 whose display content is controlled by the control unit 21, an operation unit 24 that outputs an operation signal to the control unit 21 in response to an input operation, and a communication unit 25 for communicating with external devices such as the detection unit 10 and a higher-level terminal.

[0024] In this embodiment, the moving object tracking process performed by the control unit 21 tracks the movement of a person within the monitoring area S using the detection results and the like acquired from the detection unit 10 via the communication unit 25. In this moving object tracking process, tracking position information on the tracking position of the person being tracked is associated with identification information such as an ID for distinguishing the person from other moving objects and the detection timing and registered in a tracking information list constructed in the storage unit 22. In the moving object tracking process, the movement of a person within the monitoring area S is tracked based on the information registered in the tracking information list. The control unit 21 that performs the moving object tracking process may correspond to an example of a "tracking unit".

[0025] In particular, in this embodiment, by using the tracking information list, the destination of the moving body that moves next from the previous tracking position Po is predicted for each identification information as a movement predicted position Pp. The movement predicted position information related to the movement predicted position Pp thus predicted is registered in the tracking information list in association with the identification information.

[0026] In this embodiment, the predicted movement position Pp is predicted based on a moving average vector obtained by averaging the movement vectors of the past few frames, on the premise that the tracking positions Po of the past few frames (for example, five frames) are registered in the tracking information list. The movement vector is a vector whose starting point is the previous tracking position Po and whose ending point is the current tracking position Po.

[0027] Specifically, the moving average vector is calculated using an average speed obtained by dividing the moving distance between each frame (the distance between the i-th tracking position Po and the i+1-th tracking position Po) by the elapsed time, and an angle obtained by converting each angle observation value θi into a unit vector vi=(cosθi, sinθi) on a plane to obtain a composite vector. Note that for identification information for which the tracking positions Po for the past few frames required for prediction are not registered in the tracking information list, the moving predicted position Pp is not predicted. In addition, the moving predicted position Pp is not limited to being predicted based on the moving average vector, and may be predicted using related information registered in the tracking information list.

[0028] Then, a movement prediction range E is set according to the movement prediction position information registered in the tracking information list, and an object detection position included in this movement prediction range E is a tracking position Po of the next moving object's destination, and information on the object detection position is registered in the tracking information list as tracking position information associated with the same identification information as the movement prediction position information. Note that, as described above, the control unit 21 that performs the process of registering the tracking position information in the tracking information list may correspond to an example of an "identification unit."

[0029] For example, when the kth tracking position Po and the predicted movement position Pp are registered in the tracking information list, as illustrated in FIG. 3(A), an object detection position Px within a predicted movement range E set to be centered on the predicted movement position Pp is registered in the tracking information list as the k+1th tracking position Po.

[0030] 3B, two object detection positions Px1 and Px2 may be included in one movement prediction range E. In such a case, candidate vectors Vx1 and Vx2 are calculated from the tracking position Po, which is the prediction source of the movement prediction range E, to the object detection positions Px1 and Px2, respectively, and a predicted vector Vp is calculated from the same tracking position Po to the movement prediction position Pp. Then, information on the object detection position of the candidate vector that is most similar to the predicted vector Vp among the candidate vectors is registered in the tracking information list as tracking position information.

[0031] In this embodiment, the speed (length) Lp and angle Dp of the predicted vector Vp and the speed Lx and angle Dx of each candidate vector Vx are obtained, and a similarity F is calculated for each candidate vector Vx based on the following formulas (1) to (3). Note that the coefficient α is a value between 0 and 1, which is determined depending on whether the speed or the angle is to be emphasized in the calculation of the similarity F, and is set to a larger value as the speed is emphasized. FL = 1 - (|Lx - Lp|) / Lp (1) FD = cos(|Dx-Dp|) (2) F = α × FL + (1 − α) × FD (3)

[0032] In the example of FIG. 3B, the length Lp of the predicted vector Vp is "1.00", the angle Dp is "45.0", the length Lx1 of the candidate vector Vx1 is "0.95", the angle Dx1 is "60.0", the length Lx2 of the candidate vector Vx2 is "1.10", and the angle Dx2 is "40.0". Based on the above formulas (1) to (3) and α "0.20" which is set to "low with emphasis on angle difference", the similarity F of the candidate vector Vx1 is calculated as "0.96", and the similarity F of the candidate vector Vx2 is calculated as "0.98". In this case, the candidate vector Vx2 with the largest similarity F is considered to be the candidate vector most similar to the predicted vector Vp, and information on the object detection position Px2 of this candidate vector Vx2 is registered in the tracking information list as tracking position information.

[0033] In particular, when a moving body a identified by a tracking position Poa and a moving body b identified by a tracking position Pob are temporarily close to each other, the same two object detection positions Px1 and Px2 may be included in the two predicted movement ranges Ea and Eb, respectively, as illustrated in Fig. 4. In such a case, a similarity F is calculated for each object detection position of the two moving bodies.

[0034] Specifically, a similarity F between a candidate vector Vx1a from the tracking position Poa to the object detection position Px1 and a predicted vector Vpa from the movement predicted position Ppa, and a similarity F between a candidate vector Vx2a from the tracking position Poa to the object detection position Px2 and a predicted vector Vpa from the movement predicted position Ppa are calculated. Similarly, a similarity F between a candidate vector Vx1b from the tracking position Pob to the object detection position Px1 and a predicted vector Vpb from the movement predicted position Ppb, and a similarity F between a candidate vector Vx2b from the tracking position Pob to the object detection position Px2 and a predicted vector Vpb from the movement predicted position Ppb are calculated.

[0035] Here, for example, as illustrated in FIG. 5, a case is assumed in which similarity F is calculated for each of tracking positions Poa and Pob with respect to object detection positions Px1 and Px2.

[0036] In such a case, if the object detection position Px2 is specified as the destination of the tracking position Poa based on the similarity F for the tracking position Poa previously calculated in the order of the identification numbers, only the object detection position Px1 is included within the predicted movement range Eb of the tracking position Pob, and therefore the object detection position Px1 is specified as the destination of the tracking position Pob. In such a case, the sum of the similarity F of the object detection position Px1 and the similarity F of the object detection position Px2 is 0.8 (=0.1+0.7). On the other hand, when the object detection position Px1 is specified as the destination of the tracking position Poa and the object detection position Px2 is specified as the destination of the tracking position Pob, the sum of the similarity F is 1.2 (=0.3+0.9), and the above-mentioned specification in the order of the identification numbers does not result in a combination with the maximum similarity.

[0037] For this reason, in this embodiment, as described above, when the same object detection position is included in each of the movement prediction ranges E, the object detection position to which the moving body will move is specified so as to obtain a combination with maximum similarity. Note that the method of obtaining the combination with maximum similarity is not limited to using the sum of each similarity, and the combination with maximum similarity may be obtained using, for example, the Hungarian method or the like.

[0038] The moving object tracking process performed by the control unit 21 in this embodiment will be described in detail below with reference to the flowcharts shown in FIGS. When the control unit 21 starts a moving object tracking process in response to a predetermined operation on the operation unit 24 or an instruction from a higher-level terminal via the communication unit 25, the initial setting process shown in step S101 of FIG. 6 initializes the tracking information list, etc., and makes it possible to obtain the detection result of the moving object by the detection unit 10.

[0039] Next, a detection position information acquisition process shown in step S102 is performed, and detection results for one scan (one frame) are acquired from the detection unit 10. Then, a tracking information list acquisition process shown in step S103 is performed, and processing is performed to read and acquire the tracking information list from the storage unit 22. Note that if it is immediately after the start of the moving object tracking process, a tracking information list in which tracking position information, predicted movement position information, and the like are not registered is acquired.

[0040] Next, in the determination process of step S104, it is determined whether or not there is an unsearched moving object. If there is no moving object in the monitoring area S (No in S104) because it is immediately after the start of the moving object tracking process, there is also no moving object (object detection position) to which identification information has not been assigned (No in S105), so the process from step S102 is repeated.

[0041] During the above repeated process, when a person first enters the monitoring area S where no person had been moving and an object detection position indicating the position of the person is acquired from the detection unit 10 (S102), it is determined that there is a moving object (object detection position) to which no identification information has been assigned (Yes in S105). In this case, an identification information assignment process shown in step S106 is performed, and new identification information is assigned to the moving object. Then, a tracking information registration process shown in step S107 is performed, and the object detection position is treated as a tracking position, and the tracking position information is associated with the identification information and detection timing and registered in the tracking information list.

[0042] After the tracking position information and the like are registered in the tracking information list as described above, the process from step S102 is performed again, and an object detection position that is the next destination of the moving object to which the identification information has been assigned as described above is detected, and it is determined that an unsearched moving object exists (Yes in S104). In this case, for one moving object, it is determined whether or not the object detection position exists within the predicted movement range E in the determination process of step S108 in FIG. 7. When multiple moving objects are being tracked, the process from step S108 onwards is performed for the first moving object, and then the process from step S108 onwards is performed for the second moving object after the determination of No in the above step S104, and the process from step S108 onwards is repeated according to the number of moving objects being tracked.

[0043] Here, if the tracking position information for the past few frames for which a moving average vector can be calculated is not registered in the tracking information list, the result of the determination in step S108 is Yes, on the premise that the object detection position acquired in step S102 is within a predetermined range centered on the tracking position registered in the tracking information list acquired in step S103.

[0044] In this case, since one object detection position exists within the movement prediction range E (No in S109), a tracking information update process shown in step S115 is performed. In this process, the tracking information list is updated so that information on the object detection position existing within the movement prediction range E is registered as the current tracking position information in association with the identification information together with the detection timing. Furthermore, when the tracking position information is accumulated in the tracking information list so that a moving average vector can be calculated, the tracking information list is updated so that the movement prediction position information predicted based on the moving average vector calculated as described above is registered in association with the same identification information as the above tracking position information. Note that the control unit 21 that performs the process of predicting the movement prediction position and the process of registering the movement prediction position information in the tracking information list may correspond to an example of a "prediction unit".

[0045] After the above tracking information update process, if the tracking information list has been updated for all moving objects, and there are no other unsearched moving objects (No in S104) and there are no object detection positions to which identification information has not been assigned (No in S105), the process from step S102 is repeated.

[0046] When such repeated processing results in the tracking position information for which a moving average vector can be calculated being registered in the tracking information list, in the determination processing of the above step S108, it is determined whether or not the object detection position is present within the movement prediction range E set in accordance with the tracking position information.

[0047] During the above repeated process, if another person enters the monitoring area S and there is an object detection position to which no identification information is assigned because it is not included in the movement prediction range E even though the tracking information list is updated for all moving objects (Yes in S105), identification information is assigned to the moving object (object detection position) (S106). Then, a tracking information registration process shown in step S107 is performed, and the object detection position is treated as a tracking position, and the tracking position information is associated with the identification information and detection timing and registered in the tracking information list.

[0048] In this way, when there is a plurality of pieces of identification information for moving objects to be tracked, the processes from step S108 onwards are repeated for each piece of identification information.

[0049] After that, if the identification information of the moving object to be tracked is further increased and two object detection positions are included in one movement prediction range E as shown in Fig. 3(B) (Yes in S108 and S109), a prediction vector calculation process shown in step S110 is performed. In this process, a prediction vector is calculated as described above based on the tracking position information (tracking position Po in Fig. 3(B)) and movement prediction position information (movement prediction position Pp in Fig. 3(B)) registered together with the identification information of the search target.

[0050] Next, a candidate vector calculation process is performed as shown in step S111. In this process, a candidate vector is calculated for each object detection position as described above based on each object detection position (object detection positions Px1, Px2 in FIG. 3B) and tracking position information (tracking position Po in FIG. 3B) registered together with the identification information of the search target.

[0051] Once the predictive vector and each candidate vector are calculated as described above, a destination identification process is performed as shown in step S112, and information regarding the object detection position of the candidate vector that is most similar to the predictive vector among the candidate vectors is identified as the tracking position information that will be the destination.

[0052] Then, if the same object detection position has not already been specified as tracking position information by other identification information (No in S113), the tracking position information specified as described above is associated with the identification information and the detection timing and registered in the tracking information list (S115). Note that the control unit 21, which performs the process of specifying the moving object and the process of registering the tracking position information in the tracking information list, may correspond to an example of the "specifying unit".

[0053] On the other hand, as illustrated in FIG. 4, if the same two object detection positions are included within the two predicted movement ranges, and the same object detection positions have already been identified as tracking position information by other identification information (Yes in S113), a destination re-identification process shown in step S114 is performed.

[0054] In this process, after the similarity F is calculated for each object detection position of each moving body, as described above, the object detection position to which the moving body will move is identified so as to obtain a combination with the maximum similarity. This destination re-identification process may also result in the same identification result as the destination identification process in step S112. In the example of FIG. 5, the object detection position Px1 is re-identified as the destination of the tracking position Poa, and the object detection position Px2 is re-identified as the destination of the tracking position Pob.

[0055] In the case where the tracking position information is reset as described above, the re-specified tracking position information is associated with the identification information and the detection timing and registered in the tracking information list in the tracking information update process shown in step S115. In the case where the tracking position information already specified for the other identification information needs to be changed in response to the resetting, the tracking position information is changed in this tracking information update process. In the example of FIG. 5, in the case where the tracking position information based on the object detection position Px2 specified as the destination of the tracking position Poa associated with the other identification information before the re-specification is registered in the tracking information list, the tracking position information is changed to the tracking position information based on the object detection position Px1 in response to the resetting.

[0056] On the other hand, if the object detection position is not present within the movement prediction range E (No in S108), it is determined that there is a high possibility that the moving object to be searched for has stopped temporarily. In such a case, the above-mentioned tracking information update process is not performed, and the process from step S104 is performed to identify the destination of the next unsearched moving object. In other words, if the moving object to be searched for is expected to stop temporarily, the tracking position information, etc. associated with the identification information of the moving object to be searched for is not updated.

[0057] By repeating the above-mentioned process, the tracking location information, etc. is sequentially registered in the tracking information list. Therefore, by using the tracking location information, etc. in the tracking information list, the movement path of a moving object to which identification information has been assigned can be tracked for each identification information.

[0058] As described above, in the moving object tracking system 1 according to this embodiment, tracking position information of a moving object is associated with identification information and detection timing and registered in the tracking information list. In addition, using the tracking information list, the next destination of the moving object is predicted as a predicted movement position for each identification information, and the predicted movement position information is associated with the identification information and registered in the tracking information list. Then, the moving object is specified by the object detection position being included within the predicted movement range E set according to the predicted movement position information, and information regarding the object detection position is registered in the tracking information list as tracking position information associated with the same identification information as the predicted movement position information.

[0059] This makes it possible to identify the next object detection position based on the information registered in the tracking information list, i.e., the predicted movement range E predicted from the previous tracking positions of the moving object, so that the movement path of the moving object can be tracked without obtaining characteristic information such as the color or dimensions of the detected object.

[0060] In particular, in the moving object tracking process, when multiple object detection positions are included within one movement prediction range E (Yes in S109), candidate vectors from the tracking position that is the prediction source of the movement prediction range E to the object detection positions are calculated (S110), and a predicted vector is calculated from the same tracking position to the movement prediction position (S111), and information regarding the object detection position of the candidate vector that is most similar to the predicted vector among the multiple candidate vectors is registered in the tracking information list as tracking position information.

[0061] As a result, even if multiple object detection positions are included within the movement prediction range E due to multiple moving bodies temporarily coming into close proximity, the next object detection position to be moved to can be identified by taking into account the movement direction, etc., thereby improving the tracking accuracy of the moving body.

[0062] [Second embodiment] Next, a moving object tracking system according to a second embodiment of the present invention will be described with reference to the drawings. The second embodiment is different from the first embodiment in that no identification information is added when the object detection position is detected temporarily. Therefore, the same components as those in the first embodiment are denoted by the same reference numerals and the description thereof is omitted.

[0063] In the above first embodiment, when the tracking information list is updated for all moving objects being tracked (No in S104), if there is an object detection position to which no identification information has been assigned (Yes in S105), identification information is assigned to that object detection position so that it becomes a tracking position (S106).

[0064] However, there are cases where identification information is assigned to object detection positions that are not continuously detected, for example, object detection positions that are temporarily detected as noise, etc., and there is a possibility that noise, etc. that is not originally a tracking target will be tracked. On the other hand, when a moving object cannot be detected temporarily due to the influence of a blind spot of the detection unit 10, etc., if the assignment of identification information to the moving object is cancelled, the tracking result will be lost, and therefore re-tracking will not be possible even if the moving object is detected again later.

[0065] For this reason, in this embodiment, when the tracking information list is updated for all moving bodies being tracked, if there is an object detection position that is not assigned identification information because it is not included in the predicted movement range E, identification information is assigned to the object detection position that satisfies specified conditions and it is newly registered in the tracking information list.

[0066] In this embodiment, the predetermined condition is a movement of a predetermined distance Xth and a continuation of detection a predetermined number of times (for example, two times excluding the first time). Specifically, for example, as illustrated in FIG. 8(A), when a movement (Xa1, Xa2) of a distance equal to or greater than the predetermined distance Xth is detected a predetermined number of times (twice) or more, it is determined that the predetermined condition is satisfied and the identification information is to be given. On the other hand, as illustrated in FIG. 8(B), when the distance Xb1 from the object detection position Pb1 to the object detection position Pb2 is less than the predetermined distance Xth, the number of times the object detection position Pb1 is detected is not counted, and when the distance Xb2 from the object detection position Pb2 to the object detection position Pb3 is less than the predetermined distance Xth, the number of times the object detection position Pb2 is detected is not counted, so that the predetermined condition is not satisfied and the identification information is not given. Also, as illustrated in FIG. 8(C), if object detection position Pc2 is not detected and only two object detection positions Pc1, Pc3 are detected, the object detection positions have not been detected continuously for a predetermined number of times or more, and therefore no identification information is assigned as the above-mentioned predetermined condition is not satisfied.

[0067] For this reason, in the moving object tracking process in this embodiment, information on an object detection position that is not included within the movement prediction range E is registered as tentative tracking position information in a tentative tracking information list constructed in the storage unit 22. Then, when a determination is made a predetermined number of times for the same object that the object detection position is away from the previously registered tentative tracking position by a predetermined distance, information on the object detection position is registered in the tracking information list as tracking position information in association with newly assigned identification information.

[0068] The moving object tracking process performed by the control unit 21 in this embodiment will be described in detail below with reference to the flowcharts shown in FIG. 9 and FIG. As in the first embodiment, when the tracking information list is updated for all moving objects being tracked, if there is an object detection position to which no identification information is assigned because it is not included in the movement prediction range E (Yes in S105 of FIG. 9), a determination is made in step S201 as to whether or not the object detection position has moved a predetermined distance from the previous tentative tracking position. If the object detection position is the first detection and the tentative tracking position is not registered, the determination is No in step S201, and a tentative tracking information recording process shown in step S203 is performed. In this process, information on the object detection position is registered in the tentative tracking information list as information on the tentative tracking position together with the detection timing. In this way, no identification information is assigned to the object detection position information registered in the tentative tracking information list as information on the tentative tracking position, and the process from step S102 is performed.

[0069] After that, if there is an object detection position that is not included in the movement prediction range E and therefore is not assigned identification information (Yes in S105), and the movement distance from the tentative tracking position registered in the tentative tracking information list to the object detection position is equal to or greater than the predetermined distance and equal to or less than the limit distance that the object can move, it is determined that the object has moved the predetermined distance (Yes in S201). In this case, in the determination process of step S202, it is determined whether the object has been continuously detected as a tentative tracking position a predetermined number of times or more.

[0070] Here, if the tentative tracking position was registered only last time, it is determined that the object has not been detected a predetermined number of times as a tentative tracking position (No in S202), and the tentative tracking information recording process of step S203 is performed. In this case, the information on the object detection position is registered in the tentative tracking information list as the information on the second tentative tracking position together with the detection timing.

[0071] Thereafter, when it is determined that the object detection position is a predetermined distance away from the previously registered tentative tracking position for the same object a predetermined number of times (Yes in S201 and S202), the processing from step S106 onwards is performed, and the information on the object detection position is assigned with identification information as tracking position information and registered in the tracking information list. Note that the control unit 21 that performs the determinations in steps S201 and S202 and the processing in step S106 may correspond to an example of an "identification information assigning unit."

[0072] By performing the moving object tracking process as described above, identification information is not unnecessarily assigned to object detection positions that are not a predetermined distance away from the previously registered tentative tracking position, object detection positions that are not continuously detected, and object detection positions that are temporarily detected, such as noise, etc. This makes it possible to reduce the processing load for the moving object tracking process.

[0073] [Third embodiment] Next, a moving object tracking system according to a third embodiment of the present invention will be described with reference to the drawings. The third embodiment is different from the second embodiment in that the state of the moving object being tracked is monitored to determine whether it is moving or stopped. Therefore, the same reference numerals are used for the components that are substantially the same as those in the second embodiment, and the description thereof will be omitted.

[0074] In a tracking process that tracks the path of a moving object using detection results from LiDAR or other devices, if the target remains stationary for a long period of time, the identification information of the target may be released, and even if the target begins to move again, its path may not be tracked.

[0075] Therefore, in this embodiment, in order to prevent the identification information of a moving body that is temporarily stopped from being released, each identification information is monitored to see if it is moving or stopped, and the identification information is released if it has been stopped for a certain period of time. For this reason, in the moving body tracking process performed by the control unit 21, a moving flag is set for identification information that is determined to be moving, and a stopped flag is set for identification information that is determined to be stopped, based on the relationship between the object detection position and the movement prediction range E. Specifically, a moving flag is set for identification information of a moving body whose object detection position is present within the movement prediction range E, and a stopped flag is set for identification information of a moving body whose object detection position is not present within the movement prediction range E.

[0076] The moving object tracking process performed by the control unit 21 in this embodiment will be described in detail below with reference to the flowcharts shown in FIGS. As in the first embodiment, when it is determined that an unsearched moving object exists (Yes in S104 in FIG. 10), it is determined in the determination process of step S301 in FIG. 11 whether or not a state in which a stopped flag is set for the identification information (stopped state) has continued for a certain period of time. Here, if the stopped flag is not set (No in S301) as described below, the determination process of step S303 (determination process for determining whether or not a stopped flag is set) is also determined as No, and the processes from step S108 onward are performed.

[0077] Then, when one object detection position is present within the movement prediction range E (No in S109), when the same object detection position has not already been specified as tracking position information by other identification information (No in S113), or when the above-mentioned movement destination re-identification process has been performed (S114), a tracking information update process shown in step S115a is performed. In the tracking information update process in this embodiment, in addition to updating and registering the tracking position information and movement prediction position information, etc. in the tracking information list as described above, a moving flag indicating that the object is moving and its update time (moving flag setting time) are updated and registered in the tracking information list so as to be associated with the identification information, etc.

[0078] On the other hand, if the object detection position does not exist within the movement prediction range E (No in S108), a stopped flag setting process shown in step S306 is performed. In this process, a stopped flag and its update time (stopped flag setting time) are associated with identification information for which it is determined that the object detection position does not exist within the movement prediction range E, and are updated and registered in the tracking information list. Note that the control unit 21 that performs the process of step S115a and the process of step S306 may correspond to an example of a "flag setting unit" that sets a moving flag for identification information determined to be moving and sets a stopped flag for identification information determined to be stopped.

[0079] After the above-mentioned stopping flag setting process, in the process of step S307, the number of identification information for which the stopping flag is set (hereinafter also referred to as the number of stopped moving objects removed Ns) is incremented (N=N+1) and registered in the tracking information list. Note that if the stopping flag was also set last time, the number of stopped moving objects removed Ns is not incremented. The number of stopped moving objects removed Ns is a number equivalent to the total number of moving objects that can be tracked using the tracking information list minus the number of moving objects for which the moving flag is set.

[0080] For the identification information in which the stopped flag is set as described above, at the next processing of step S102 or later, the result of the judgment process of step S303 is judged as Yes, and in the judgment process of step S304, it is determined whether or not the object detection position is present within a certain distance range centered on the previous tracking position in the tracking information list.

[0081] Here, if the object detection position is within the range (e.g., about ±50%) of the certain distance (e.g., the movement speed of a general person × the time of one frame), it is determined as Yes in step S304, and a movement restart candidate list registration process shown in step S305 is performed. In this process, since there is a possibility that the object is moving differently from the prediction, information on the object detection position included within the certain distance is associated with the identification information of the previous tracking position, and is registered in the movement restart candidate list constructed in the storage unit 22 as detection position information of the movement restart candidate. Thereafter, the processes from step S108 onwards are performed.

[0082] On the other hand, if the object detection position does not exist within the range of the certain distance, it is determined as No in step S304, and the processing from step S108 onwards is performed without registering the object detection position in the movement restart candidate list. For example, if the object detection position exists in a position that is too close to the previous tracking position and therefore does not exist within the range of the certain distance (No in S303), the processing from step S108 onwards is performed without registering the object detection position in the movement restart candidate list.

[0083] Thereafter, when a tracking information update process (S115a) is performed so that the moving flag is set, or when the stopped moving object deletion count Ns is incremented after the stopped flag is set (S307), a determination process in step S308 is performed to determine whether or not the information of the movement restart candidate registered in the movement restart candidate list has been added to the tracking information list. Here, if the information of the movement restart candidate has been registered as described above (Yes in S308), the information of the movement restart candidate is deleted from the movement restart candidate list in a movement restart candidate information deletion process in step S309.

[0084] When the information of the movement restart candidate is not registered in the tracking information list (No in S308), or when the information of the movement restart candidate is deleted from the movement restart candidate list as described above (S309), an additional information update process shown in step S310 is performed. In this process, the tracking information list is updated so that the detected position of the movement restart candidate is registered as the current tracking position information together with the detection timing in association with the identification information. Furthermore, the tracking information list is updated so that the movement predicted position information predicted as described above using the information registered in the tracking information list in association with the identification information is registered in association with the same identification information as the above tracking position information. Furthermore, the moving flag indicating that the movement is in progress and its update time (moving flag setting time) are updated and registered in the tracking information list so as to be associated with the identification information, etc. Thereafter, the information registered in the movement restart candidate list is deleted in the moving restart candidate information deletion shown in step S311, and then the process from step S104 in FIG. 10 described above is performed.

[0085] Furthermore, when the information of the object detection position to which new identification information has been added is registered as tracking position information in the tracking information list (S107), in the judgment process of step S315, it is judged whether or not the number of stopped moving objects to be deleted Ns is greater than 0. Here, since there is identification information with a stopped flag set, if the number of stopped moving objects to be deleted Ns is greater than 0 (Yes in S315), the tracking information list is updated so that the number of stopped moving objects to be deleted Ns is decremented (Ns=Ns-1), and the process from step S102 is performed. On the other hand, if the number of stopped moving objects to be deleted Ns is 0 (No in S315), the number of stopped moving objects to be deleted Ns is not decremented, and the process from step S102 is performed.

[0086] Then, when it is determined that the stopped flag has been set a predetermined number of times in succession based on information such as the stop flag setting time in the tracking information list and that the stopped flag has been set for a certain period of time, the determination result in the determination process in step S301 is Yes, and a tracking information deletion process shown in step S302 is performed. In this process, the moving object to which the identification information of the stopped flag was assigned is deemed to be untraceable (lost), and the identification information is deleted from the tracking information list together with associated information. The control unit 21 that performs the tracking information deletion process may correspond to an example of a "deletion unit".

[0087] On the other hand, if the moving object to which the identification information of the stop flag is assigned starts moving again before the fixed period of time has elapsed, the stop flag is set (Yes in S303) and the object detection position is within the fixed distance range (Yes in S304), so the movement restart candidate list registration process of step S305 is performed. After that, since the object detection position is within the movement prediction range E (Yes in S108), a tracking information update process (S115a) is performed so that the moving flag is set.

[0088] As described above, in the moving object tracking system 1 according to this embodiment, a moving flag is set for identification information determined to be moving based on the relationship between the object detection position and the predicted movement range E (S115a), and a stopped flag is set for identification information determined to be stopped (S306). Then, identification information for which a stopped flag has been set for a certain period of time is deleted from the tracking information list together with associated information (S302).

[0089] As a result, if a moving object becomes untraceable because it has moved into a blind spot of the detection unit 10, the stopped flag is set for a certain period of time, and the identification information of the moving object is deleted from the tracking information list. Therefore, there is no unnecessary movement prediction for the identification information of an untraceable moving object, and the processing load for tracking the moving object can be reduced. In addition, if a moving object that has been temporarily stopped moves again, the flag of the moving object is reset to a moving flag, and tracking processing can be continued for the moving object.

[0090] In particular, in this embodiment, by using the number of removed stopped moving objects Ns counted as described above, the total number Nt of moving objects can be accurately calculated even when the moving objects that have entered the monitoring area S temporarily stop and then move again. That is, the number of identification information items for which a stop flag has been set as a temporary stop because the object detection position is not present within the predicted movement range E is recorded as the number of removed stopped moving objects Ns using the tracking information list, and the number of removed stopped moving objects Ns is reduced when new identification information is assigned in the above step S107.

[0091] If the number of stopped moving objects removed Ns is not taken into consideration, for example, if three moving objects are being tracked and the total number Nt of moving objects that have entered the monitoring area S is calculated as "3," if one moving object temporarily stops and then starts moving again, the total number Nt may be erroneously calculated as "4" as a new moving object has been detected.

[0092] For this reason, as described above, the number of identification information with the stop flag set is recorded as the number of stopped moving objects removed Ns, and when new identification information is assigned, the number of stopped moving objects removed Ns is reduced. As a result, in the above example, even if one moving object moves again, the number of stopped moving objects removed Ns is counted because the moving object was temporarily stopped, so by taking this number of stopped moving objects removed Ns into consideration, the total number can be correctly calculated as "3". Note that the control unit 21 that performs a process of calculating the total number Nt of moving objects that have entered the monitoring area S using the number of newly assigned identification information and the number of stopped moving object removed Ns, which is the number of identification information with the stop flag set, can correspond to an example of a "total number calculation unit".

[0093] The present invention is not limited to the above-described embodiment, and may be embodied as follows, for example. (1) The movement prediction range E is not limited to being set to a range whose outer edge is a perfect circle centered on the movement prediction position Pp as illustrated in Figures 3 and 4, but may be set to, for example, a range whose outer edge is an ellipse centered on the movement prediction position Pp.

[0094] (2) The present invention is not limited to being adopted in a moving object tracking system whose monitoring area S is a room with three entrances Sa, Sb, and Sc, but may be adopted in a moving object tracking system whose monitoring area is a room with one or two entrances, or may be adopted in a moving object tracking system whose monitoring area is a room with four or more entrances. Furthermore, the present invention is not limited to being adopted in a moving object tracking system whose monitoring area is a room with entrances, but may be adopted in a moving object tracking system whose monitoring area is, for example, an area in front of a store, a public facility, an event venue, or the like where many people move, or may be adopted in a moving object tracking system whose tracking target is an object other than a person.

[0095] (3) The detection unit 10 is not limited to being configured as a LiDAR, and may be configured as a two-dimensional scanning distance measuring device that employs other optical measurement methods.

[0096] 1. Mobile tracking system 10. Detection section 20 Tracking Device 21...Control unit (prediction unit, identification unit, tracking unit, identification information assignment unit, flag setting unit, deletion unit) E. Movement prediction range Po Tracking Position Pp predicted movement position Px Object detection position Vp Prediction Vector Vx candidate vector S Surveillance Area

Claims

1. A moving object tracking system for tracking a moving object that has entered a predetermined monitoring area, comprising: a detection unit that detects a position of an object, including a moving object, within the predetermined monitoring area by receiving reflected light from the object in response to light projected toward the predetermined monitoring area, as an object detection position; a storage unit having a tracking information list in which tracking position information regarding the tracking position of a moving object is registered in association with identification information for distinguishing the moving object from other moving objects and detection timing; a prediction unit that uses the tracking information list to predict a next destination of the moving object as a movement predicted position for each of the identification information, and registers movement predicted position information relating to the movement predicted position in the tracking information list in association with the identification information; an identification unit that identifies a moving object when the object detection position is included in a movement prediction range that is set according to the movement prediction position information, and registers information regarding the object detection position in the tracking information list as the tracking position information associated with the same identification information as the movement prediction position information; a tracking unit that tracks the movement of a moving object within the predetermined monitoring area based on the information registered in the tracking information list; A moving object tracking system comprising:

2. The moving object tracking system of claim 1, characterized in that, when a plurality of the object detection positions are included within one of the movement prediction ranges, the identification unit calculates candidate vectors from the tracking position that is the prediction source of the movement prediction range to the object detection position as a starting point, and calculates a predictive vector from the same tracking position to the movement prediction position as a starting point, and registers information regarding the object detection position of a candidate vector that is most similar to the predictive vector among the plurality of candidate vectors in the tracking information list as the tracking position information.

3. the storage unit has a tentative tracking information list in which the object detection position that is not included in the movement prediction range is registered as a tentative tracking position; The moving object tracking system according to claim 1, further comprising an identification information assignment unit that, when a determination is made a predetermined number of times for the same object that the object detection position is away from a previously registered tentative tracking position by a predetermined distance, associates information regarding the object detection position with the newly assigned identification information as the tracking position information and registers the information in the tracking information list.

4. a flag setting unit for setting a moving flag for the identification information determined to be moving and setting a stopped flag for the identification information determined to be stopped based on a relationship between the object detection position and the predicted movement range; a deletion unit that deletes, from the tracking information list, the identification information for which the stop flag has been set for a certain period of time together with associated information; The moving object tracking system according to claim 1, further comprising:

5. a flag setting unit for setting a moving flag for the identification information determined to be moving and setting a stopped flag for the identification information determined to be stopped based on a relationship between the object detection position and the predicted movement range; a deletion unit that deletes, from the tracking information list, the identification information for which the stop flag has been set for a certain period of time together with associated information; a total number calculation unit that calculates the total number of the moving objects that have entered the specified monitoring area by using the number of the identification information newly assigned by the identification information assignment unit and the number of the identification information for which the stopped flag has been set by the flag setting unit; 4. The moving object tracking system according to claim 3, further comprising:

Images