MOBILE OBJECT TRACKING DEVICE, METHOD, AND PROGRAM

Abstract

The present invention enables accurate tracking of a mobile object across time-series images. A detection means (11) detects a mobile object from each of time-series images obtained by capturing an image of a road. A prediction means (12) uses past information indicating positions on the road where a mobile object was detected in the past to predict a region to which the mobile object is to move. With regard to a region to which a mobile object detected from a first image is predicted to move, if a mobile object is detected in that region from a second image, then a tracking means (13) treats the mobile object detected from the first image and the mobile object detected from the second image as the same mobile object and tracks same.

Description

Mobile object tracking device, method, and computer-readable medium

[0001] The present disclosure relates to a mobile object tracking device, method, and computer-readable medium.

[0002] As a related technique, Patent Document 1 discloses a moving object tracking device that tracks a moving object included in multiple images captured in time series. The moving object tracking device acquires paired features of the tracked vehicle from the t-th image captured by a camera, and then searches for a destination area of ​​the tracked vehicle in the t+1-th image captured by the camera. In the process of searching for the destination area, the moving object tracking device extracts a number of image areas that are candidate destination areas from the t+1-th image. The candidate destination image areas can be determined by predicting the vehicle's movement direction and movement speed from previous vehicle tracking results.

[0003] The moving object tracking device searches for a destination area of ​​the vehicle by searching for a destination candidate that is most similar to the positive sample of the tth image among multiple destination candidates extracted from the t+1th image based on paired features. Specifically, the moving object tracking device extracts pixel pairs from the same positions as the multiple pixel pairs extracted as paired features of the positive sample in each image area of ​​the destination candidate. The moving object tracking device calculates the similarity between the positive sample and the destination candidate using the paired features (pixel pairs) of the positive sample and the pixel pairs extracted from the destination candidate. The moving object tracking device calculates the similarity between each of the multiple destination candidates extracted from the t+1th image and the positive sample, and determines the destination candidate with the greatest similarity as the final destination of the tracked vehicle.

[0004] JP 2011-118450 A

[0005] In Patent Document 1, a vehicle is tracked by searching for areas with similar paired features between time-series images. However, when the feature amount is small, it is possible to detect a vehicle from an image, but it is difficult to track the vehicle between time-series images. In particular, when the camera frame rate is low, the amount of movement of the vehicle between time-series images is large, making it difficult to track the vehicle between time-series images.

[0006] In Patent Document 1, a moving object tracking device extracts a large number of destination candidates in the t+1th image and determines the destination of the tracked vehicle based on paired features. Regarding the extraction of destination candidates, Patent Document 1 describes predicting the vehicle's movement direction and movement speed from previous vehicle tracking results. However, since Patent Document 1 uses the previous vehicle tracking results to extract destination candidates, there is a problem in that the accuracy of vehicle tracking decreases in tracking difficult situations.

[0007] In view of the above circumstances, an object of the present disclosure is to provide a moving object tracking device, method, and computer-readable medium that can accurately track a moving object between time-series images.

[0008] To achieve the above object, the present disclosure provides, as a first aspect, a moving object tracking device, which includes: a detection means for detecting a moving object from each of time-series images of a road; a prediction means for predicting a destination area of ​​the moving object using past information indicating past detection positions of the moving object on the road; and a tracking means for tracking the moving object detected from the first image and the moving object detected from the second image as the same moving object, when a moving object is detected from a second image captured at a time later than the time the first image was captured in the predicted destination area for the moving object detected from a first image included in the time-series images.

[0009] The present disclosure provides, as a second aspect, a moving object tracking method, which includes: detecting a moving object from a first image included in a time-series image of a road; predicting a destination area of ​​the moving object detected from the first image using past information indicating past detection positions of the moving object on the road; detecting the moving object from a second image included in the time-series image that was captured at a time later than the time the first image was captured; and, if the moving object detected from the second image is detected in the predicted destination area for the moving object detected in the first image, tracking the moving object detected from the first image and the moving object detected from the second image as the same moving object.

[0010] In a third aspect, the present disclosure provides a computer-readable medium storing a program for causing a computer to execute a process including: detecting a moving object from a first image included in a time-series image of a road; predicting a destination area of ​​the moving object detected from the first image using past information indicating past detection positions of the moving object on the road; detecting the moving object from a second image included in the time-series image that was captured at a time later than the time the first image was captured; and, if the moving object detected from the second image is detected in the predicted destination area for the moving object detected in the first image, tracking the moving object detected from the first image and the moving object detected from the second image as the same moving object.

[0011] The moving object tracking device, method, and computer-readable medium according to the present disclosure can accurately track a moving object between time-series images.

[0012] 1 is a block diagram showing a schematic configuration example of a moving object tracking device according to the present disclosure. 2 is a block diagram showing a moving object tracking device according to an embodiment of the present disclosure. 3 is a flowchart showing an operation procedure in the moving object tracking device. 4 is a schematic diagram showing a state of an intersection at time t. 5 is a schematic diagram showing a state of an intersection at time t+1. 6 is a schematic diagram showing a state of an intersection at time t+2. 7 is a schematic diagram showing a state of an intersection in a certain phase. 8 is a block diagram showing a configuration example of a computer device.

[0013] Prior to describing embodiments of the present disclosure, an overview of the present disclosure will be described. FIG. 1 shows a schematic configuration example of a moving object tracking device according to the present disclosure. The moving object tracking device 10 includes a detection unit 11, a prediction unit 12, and a tracking unit 13. The detection unit 11 detects a moving object from each of time-series images of a road. Here, the time-series images refer to, for example, two or more images captured consecutively in time using the same imaging device. The time-series images include a first image and a second image captured at a time later than the time at which the first image was captured.

[0014] The prediction means 12 predicts a destination area of ​​the detected moving object using past information indicating past detected positions of the moving object on the road. When a moving object is detected in the second image in the predicted destination area of ​​the moving object detected in the first image, the tracking means 13 tracks the moving object detected in the first image and the moving object detected in the second image as the same moving object. Here, tracking means, for example, associating moving objects that appear in images taken at different times as the same moving object.

[0015] In the present disclosure, the prediction means 12 predicts a destination area of ​​the moving object detected in the first image using a previously detected position of the moving object. When a moving object is detected in the second image within the predicted destination area, the tracking means 13 tracks the moving object detected in the first image and the moving object detected in the second image as the same moving object. In the present disclosure, an area including a position where a moving object has been detected on a road in the past and is therefore likely to pass through can be predicted as the destination area. Therefore, the moving object tracking device according to the present disclosure can accurately track the moving object between time-series images.

[0016] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Note that the following description and drawings have been omitted and simplified as appropriate for clarity of explanation. In addition, in the following drawings, the same or similar elements are designated by the same reference numerals, and duplicate explanations are omitted as necessary.

[0017] 2 shows a moving object tracking device according to an embodiment of the present disclosure. The moving object tracking device 100 includes an image acquisition unit 101, a detection unit 102, a prediction unit 103, a tracking unit 104, and a detected position storage unit 105. The moving object tracking device 100 may be configured using, for example, a computer including at least one processor and at least one memory. At least a portion of the functions of each unit of the moving object tracking device 100 may be realized by the processor operating in accordance with a program read from the memory.

[0018] The image acquisition unit 101 acquires time-series images from, for example, one or more cameras 210. The cameras 210 capture images of an area including a road. The cameras 210 are installed on road facilities such as traffic lights. The image acquisition unit 101 acquires time-series images from the cameras 210 via a network. The network includes, for example, a network using a communication line standard such as LTE (Long Term Evolution). The network may include a wireless communication network such as Wi-Fi (registered trademark) or a fifth-generation mobile communication system.

[0019] The moving object tracking device 100 may be disposed at each intersection, for example. Alternatively, one moving object tracking device 100 may be disposed corresponding to a predetermined geographical area, and the moving object tracking device 100 may receive time-series images from cameras 210 installed within the predetermined geographical area. The image acquisition unit 101 may acquire, as time-series images, three-dimensional point cloud data (three-dimensional point cloud images) acquired using, for example, LiDAR (light detection and ranging). The time-series images include, for example, multiple images of an intersection including a road, captured in time series. The time-series images include a first image and a second image. The second image is assumed to be an image captured at a time later than the time the first image was captured.

[0020] The detection unit 102 detects a moving object from the time-series images acquired by the image acquisition unit 101. For example, the detection unit 102 detects the area of ​​the moving object included in the image as the position of the moving object. The method used to detect the moving object is not limited to a particular method. The detection unit 102 can detect the position of the moving object using a known algorithm. When multiple moving objects are included in the image, the detection unit 102 detects the positions of each of the multiple moving objects. The detection unit 102 may correct distortion of the image and detect the absolute position, i.e., the position of the moving object in real space. The detection unit 102 may extract feature values ​​from the image for the detected moving object. The detection unit 102 corresponds to the detection means 11 shown in FIG. 1.

[0021] The detection unit 102 may identify the type of the detected moving object. The type of moving object may include, for example, a private car, a bus, a truck, a motorcycle, a bicycle, a person, and a streetcar. The type of moving object may be broadly categorized, for example, into four-wheeled vehicles and two-wheeled vehicles. In this case, four-wheeled vehicles may be classified into large vehicles and standard-sized or compact vehicles. The detection unit 102 may analyze, for example, the shape, size, color, and license plate information of the moving object, and identify or estimate the type of each detected moving object. The detection unit 102 may detect the moving object and identify its type, for example, by applying an image to an AI (Artificial Intelligence) model.

[0022] The detection unit 102 stores the position of the detected moving object in the detection position storage unit 105. The detection position storage unit 105 stores or accumulates the position of the detected moving object, i.e., the detected position of the moving object, as past information. The detection position storage unit 105 may be configured using a storage device such as a hard disk drive or an SSD (Solid State Drive). The detection position storage unit 105 may store the detected position of the moving object for each type of moving object. In other words, the detection position storage unit 105 may store the detected position of the moving object and the identified type of moving object in association with each other. Note that the detection position storage unit 105 does not necessarily have to be included in the moving object tracking device 100. For example, the detection position storage unit 105 may be configured as an external storage connected to the moving object tracking device 100 via a network.

[0023] The prediction unit 103 acquires past information, i.e., data on past detection positions of moving objects at intersections or roads, from the detection position storage unit 105. The prediction unit 103 uses the acquired past information to predict a destination area in an image at a later time of the moving object detected by the detection unit 102. The prediction unit 103 predicts, for example, an area ahead in the traveling direction of the moving object and including a position where the moving object has been detected in the past, as the destination area.

[0024] For example, an image captured at time t is defined as the first image, and an image captured at a time later than time t, for example, time t+1, is defined as the second image. The prediction unit 103 uses data on past detection positions of moving objects to predict the position range in the second image of the moving object detected in the first image, i.e., the destination area. The prediction unit 103 predicts, for example, the position of the moving object in the second image. The prediction unit 103 predicts, as the destination area of ​​the moving object in the second image, an area that is a region that includes a predetermined margin added to the predicted position and that includes a position where a moving object has been detected in the past. The prediction unit 103 corresponds to the prediction means 12 shown in FIG. 1 .

[0025] The prediction unit 103 may predict multiple directions in which the moving body may travel based on the structure of the intersection and the position of the moving body, and may determine multiple predicted positions using the predicted traveling directions. For example, when the moving body is likely to turn right or go straight at the intersection, the prediction unit 103 may predict a predicted position of the moving body when turning right and a predicted position of the moving body when going straight. In this case, the prediction unit 103 may merge a destination area when turning right and a destination area when going straight, and predict the merged area as a destination area in the second image.

[0026] When the detection position storage unit 105 stores past information for each type of moving object, i.e., past detection positions of the moving object, the prediction unit 103 may acquire the past detection positions corresponding to the type of detected moving object and predict the destination area of ​​the moving object in the second image. For example, large vehicles and standard-sized vehicles may differ in the way they navigate through intersections, which may result in different detection positions. Furthermore, four-wheeled vehicles and two-wheeled vehicles may differ in the way they navigate through intersections, which may result in different detection positions. Therefore, it is believed that prediction accuracy can be improved by predicting the destination area using past detection positions corresponding to the type of moving object.

[0027] The prediction unit 103 may acquire the light status of a traffic light installed at the intersection and predict a destination area of ​​the moving object in the second image based on the acquired light status. The prediction unit 103 may acquire the light status of the traffic light from, for example, a traffic light control panel. The prediction unit 103 may analyze a camera image to acquire the light status. For example, if the light status of the traffic light indicates that proceeding is prohibited, the prediction unit 103 may predict that the moving object will stop before a stop line and predict the destination area based on the prediction. Furthermore, if the light status of the traffic light indicates that proceeding is permitted only in a specific direction, the prediction unit 103 may predict that the moving object will proceed in that specific direction and predict the destination area based on the prediction.

[0028] The tracking unit 104 tracks the moving object detected between the time-series images based on the position of the moving object detected by the detection unit 102 and the destination area of ​​the moving object predicted by the prediction unit 103. The tracking unit 104 determines whether or not a moving object has been detected in the second image in the destination area predicted by the prediction unit 103 for the moving object detected in the first image. If a moving object has been detected in the predicted destination area, the tracking unit 104 tracks the moving object detected in the first image and the moving object detected in the second image as the same moving object.

[0029] The tracking unit 104 may calculate a similarity between a feature amount of the moving object detected in the first image and a feature amount of the moving object detected in the second image. If the similarity between the feature amounts is equal to or greater than a predetermined value, the tracking unit 104 may determine that the moving object detected in the first image and the moving object detected in the second image are the same moving object. The tracking result of the tracking unit 104 can be used for purposes such as traffic volume surveys and counting the number of passing vehicles by direction. The tracking unit 104 corresponds to the tracking means 13 shown in FIG. 1.

[0030] Here, if the tracking unit 104 has already tracked a moving object at a time prior to the first image, the prediction unit 103 may predict the position of the moving object in the second image using the tracking results of the moving object. For example, the prediction unit 103 may calculate the moving speed and moving direction of the moving object from the tracking results of the past few frames, and determine the predicted position of the moving object in the second image based on the calculated moving speed and moving direction. The moving speed can be calculated, for example, from the frame rate, i.e., the time interval between time-series images, and the displacement or movement amount of the moving object.

[0031] Next, the operation procedure will be described. FIG. 3 shows the operation procedure in the moving object tracking device 100. The operation procedure in the moving object tracking device 100 is also called a moving object tracking method. The camera 210 captures an image of a road at an intersection. The image acquisition unit 101 acquires an image from the camera 210. The detection unit 102 detects a moving object from the acquired image (step S1). In step S1, the detection unit 102 may estimate or identify the type of the moving object. The prediction unit 103 acquires past information from the detection position storage unit 105 (step S2). If the type of the moving object is estimated or identified in step S1, the prediction unit 103 may acquire past information corresponding to the estimated or identified type in step S3.

[0032] The prediction unit 103 predicts the destination region of the moving object detected in step S1 in the next image using the past information acquired in step S2 (step S3). In step S3, the prediction unit 103 predicts the position of the moving object in the next image, for example, using the position of the moving object detected in step S1 as a base point. The prediction unit 103 predicts, as the destination region, a region that includes a margin added to the predicted position and that also includes a position where a moving object has been detected in the past.

[0033] The tracking unit 104 compares the position of the moving object detected in step S1 with the predicted destination area of ​​the moving object detected in a previous image, for example, the image one time before. The tracking unit 104 determines whether the moving object has been detected in the predicted destination area. If the moving object has been detected in the predicted destination area, the tracking unit 104 detects the moving object detected in step S1 and the moving object detected in the image one time before as the same moving object (step S5). If the moving object has not been detected in the predicted destination area, the tracking unit 104 determines that the moving object detected in step S1 and the moving object detected in the image one time before are different moving objects.

[0034] A specific example will be described below. FIG. 4 schematically shows the state of an intersection at time t. A moving object, vehicle 310, is about to enter the intersection. In FIG. 4, the past detected positions of the moving object stored in the detected position storage unit 105 (see FIG. 1) are represented by black circles. Note that, although not shown in FIG. 4 for the sake of simplicity, the detected position storage unit 105 also stores the past detected positions of the moving object for lanes opposite the lane in which vehicle 310 is traveling and roads intersecting the road in which vehicle 310 is traveling.

[0035] At time t, the detection unit 102 detects a vehicle 310. It is assumed that at time t-1, the vehicle 310 is detected at a detection position 320 indicated by a dashed line. The tracking unit 104 assumes that the vehicle 310 detected at time t and the vehicle detected at detection position 320 at time t-1 are tracked as the same vehicle. In this case, the prediction unit 103 predicts the position of the vehicle 310 at time t+1 based on the detection position of the vehicle 310 at time t and the detection position 320 of the vehicle 310 in the image at time t-1. The prediction unit 103 predicts, as the destination area of ​​the vehicle 310 at time t+1, an area 330 that includes the predicted position and also includes positions where a moving object has been detected in the past.

[0036] 5 schematically shows the state of an intersection at time t+1. The detection unit 102 detects a vehicle 310 from an image taken at time t+1. If the vehicle 310 is detected at time t+1 in the area 330 predicted at time t (see FIG. 4 ), the tracking unit 104 tracks the vehicle detected at time t and the vehicle detected at time t+1 as the same vehicle.

[0037] At time t+1, vehicle 310 has entered nearly halfway through the intersection, and it can be predicted that vehicle 310 will either turn right or go straight through the intersection. Based on the detected position of vehicle 310 at time t+1 and detected position 320 of vehicle 310 at time t, prediction unit 103 predicts the position of vehicle 310 at time t+2 for each of the cases of turning right and going straight. For each of the cases of turning right and going straight, prediction unit 103 predicts, as the destination area of ​​vehicle 310 at time t+1, an area that includes the predicted position and also includes a position where the moving object has been detected in the past. Prediction unit 103 predicts, as the destination area of ​​vehicle 310 at time t+2, an area 340 that is a merge of the destination area for turning right and the destination area for going straight.

[0038] 6 schematically shows the state of an intersection at time t+2. The detection unit 102 detects a vehicle 310 from an image taken at time t+2. If the vehicle 310 is detected at time t+2 in the area 340 (see FIG. 5 ) predicted at time t+1, the tracking unit 104 tracks the vehicle detected at time t+1 and the vehicle detected at time t+2 as the same vehicle.

[0039] At time t+2, vehicle 310 has changed direction, and it can be predicted that vehicle 310 will turn right at the intersection instead of going straight. Prediction unit 103 predicts the position of vehicle 310 at time t+3 based on the detected position of vehicle 310 at time t+2 and detected position 320 of vehicle 310 at time t+1. Prediction unit 103 predicts, as the destination area of ​​vehicle 310 at time t+3, an area 350 that includes the predicted position and also includes positions where a moving object has been detected in the past. If vehicle 310 is detected at time t+3 in area 350 predicted at time t+2, tracking unit 104 tracks the vehicle detected at time t+2 and the vehicle detected at time t+3 as the same vehicle.

[0040] FIG. 7 schematically shows the state of an intersection in a certain situation. Here, it is assumed that the detection position storage unit 105 stores the past detection positions of four-wheeled vehicles and the past detection positions of two-wheeled vehicles. In FIG. 7, the positions where four-wheeled vehicles have been detected in the past are represented by black circles, and the positions where two-wheeled vehicles have been detected in the past are represented by white circles. As shown in FIG. 7, four-wheeled vehicles and two-wheeled vehicles may pass through the intersection at different locations.

[0041] The detection unit 102 detects a vehicle 310, which is a four-wheeled vehicle, and a motorcycle 410, which is a two-wheeled vehicle. The prediction unit 103 predicts the positions of the vehicle 310 and the motorcycle 410 at the next time. Here, it is assumed that the prediction unit 103 predicts that the vehicle 310 and the motorcycle 410 will turn right at an intersection. The prediction unit 103 predicts, as a destination area for the vehicle 310, an area 360 that includes the predicted position of the vehicle 310 and also includes positions where a four-wheeled vehicle has been detected in the past. On the other hand, the prediction unit 103 predicts, as a destination area for the motorcycle 410, an area 420 that includes the predicted position of the motorcycle 410 and also includes positions where a two-wheeled vehicle has been detected in the past.

[0042] In the above case, if vehicle 310 is detected in predicted area 360 at the next time, tracking unit 104 tracks the vehicle detected at the previous time and the vehicle detected at the next time as the same vehicle. Also, if motorcycle 410 is detected in predicted area 420 at the next time, tracking unit 104 tracks the motorcycle detected at the previous time and the motorcycle detected at the next time as the same motorcycle. In this way, by predicting the destination area according to type, it becomes easier to track different types of moving objects when the locations at which they pass through intersections differ depending on the type.

[0043] In this embodiment, the detection position storage unit 105 stores past detection positions of the moving object. The prediction unit 103 predicts a destination area of ​​the moving object detected in the first image using the detection positions of the moving object stored in the detection position storage unit 105. When a moving object is detected in the second image in the predicted destination area, the tracking unit 104 tracks the moving object detected in the first image and the moving object detected in the second image as the same moving object. In this embodiment, the prediction unit 103 can predict, as the destination area, an area where a moving object has been detected in the past and therefore includes a position where the moving object is likely to pass. Therefore, the moving object tracking device 100 according to this embodiment can accurately track the moving object in the first image and the second image.

[0044] In the present disclosure, the moving object tracking device 100 may be configured as a computer device or a server device. Fig. 8 shows an example configuration of a computer device that may be used as the moving object tracking device 100. The computer device 500 includes a control unit (CPU: Central Processing Unit) 510, a storage unit 520, a ROM (Read Only Memory) 530, a RAM (Random Access Memory) 540, a communication interface (IF: Interface) 550, and a user interface 560.

[0045] The communication interface 550 is an interface for connecting the computer device 500 to a communication network via wired communication means, wireless communication means, etc. The user interface 560 includes a display unit such as a display, and an input unit such as a keyboard, a mouse, and a touch panel.

[0046] The storage unit 520 is an auxiliary storage device that can store various types of data. The storage unit 520 does not necessarily have to be a part of the computer device 500, but may be an external storage device or cloud storage connected to the computer device 500 via a network.

[0047] The ROM 530 is a non-volatile storage device. For example, a semiconductor storage device with a relatively small capacity, such as a flash memory, is used for the ROM 530. The programs executed by the CPU 510 can be stored in the storage unit 520 or the ROM 530. The storage unit 520 or the ROM 530 stores various programs for realizing the functions of each unit in the moving object tracking device 100, for example.

[0048] The above program can be stored and supplied to the computer device 500 using various types of non-transitory computer-readable media. Non-transitory computer-readable media include various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic recording media such as flexible disks, magnetic tapes, or hard disks; magneto-optical recording media such as magneto-optical disks; optical disk media such as compact discs (CDs) or digital versatile disks (DVDs); and semiconductor memories such as mask ROMs, programmable ROMs (PROMs), erasable PROMs (EPROMs), flash ROMs, or RAMs. The program may also be supplied to the computer using various types of temporary computer-readable media. Examples of temporary computer-readable media include electrical signals, optical signals, and electromagnetic waves. The temporary computer-readable media can be supplied to the computer via wired communication paths such as electric wires and optical fibers, or via wireless communication paths.

[0049] The RAM 540 is a volatile storage device. Various semiconductor memory devices such as a dynamic random access memory (DRAM) or a static random access memory (SRAM) are used for the RAM 540. The RAM 540 can be used as an internal buffer for temporarily storing data and the like. The CPU 510 loads a program stored in the storage unit 520 or the ROM 530 into the RAM 540 and executes it. The CPU 510 executes the program, thereby realizing the functions of each unit in the moving object tracking device 100. The CPU 510 may have an internal buffer for temporarily storing data and the like.

[0050] Although the embodiments of the present disclosure have been described in detail above, the present disclosure is not limited to the above-described embodiments, and changes and modifications to the above-described embodiments without departing from the spirit of the present disclosure are also included in the present disclosure.

[0051] 10: Mobile object tracking device 11: Detection means 12: Prediction means 13: Tracking means 100: Mobile object tracking device 101: Image acquisition unit 102: Detection unit 103: Prediction unit 104: Tracking unit 105: Detection position memory unit 210: Camera 310: Vehicle 410: Motorcycle 500: Computer device 510: Control unit 520: Memory unit 530: ROM 540: RAM 550: Communication interface 560: User interface

Claims

1. A detection means for detecting a moving object from each of time-series images of a road; a prediction means for predicting a destination area of ​​a moving object by using past information indicating a past detected position of the moving object on the road; a tracking means for tracking the moving object detected from the first image and the moving object detected from the second image as the same moving object when a moving object is detected from a second image taken at a later time than the time the first image was taken in a predicted destination area of ​​the moving object detected from a first image included in the time series images.

2. The moving body tracking device according to claim 1, wherein the prediction means predicts the position of the moving body at the time the second image is taken, and predicts as the destination area an area that includes the predicted position and also includes the detected position of the moving body in the past information.

3. 3. The moving object tracking device according to claim 2, wherein, when a moving object detected in the first image is a moving object being tracked by the tracking means, the prediction means calculates a moving speed and a moving direction of the moving object using a tracking result of the tracking means, and predicts a position of the moving object at the time when the second image is captured using the calculated moving speed and moving direction.

4. the past information is stored for each type of the moving object, The moving object tracking device according to claim 1 , wherein the prediction means predicts the destination area by using the past information corresponding to a type of the detected moving object.

5. The moving object tracking device according to claim 1 , wherein the time-series images include a plurality of images taken in time series of an intersection including the road.

6. 6. The moving object tracking device according to claim 5, wherein the prediction means further acquires a lighting state of a traffic light installed at the intersection, and predicts the destination area based on the acquired lighting state.

7. The detection means extracts a feature amount of the detected moving object from the time-series images; 4. The moving object tracking device according to claim 1, wherein the tracking means calculates a similarity between the features of the moving object detected in the first image and the features of the moving object detected in the second image, and if the calculated similarity is equal to or greater than a predetermined value, tracks the moving object detected from the first image and the moving object detected from the second image as the same moving object.

8. Detecting a moving object from a first image included in a time series of images of a road; predicting a destination area of ​​the moving object detected from the first image using past information indicating a past detection position of the moving object on the road; detecting a moving object from a second image included in the time series images and captured at a time later than the time when the first image was captured; A moving object tracking method comprising: when a moving object detected from the second image is detected in the area of ​​the predicted destination of the moving object detected from the first image, tracking the moving object detected from the first image and the moving object detected from the second image as the same moving object.

9. Detecting a moving object from a first image included in a time series of images of a road; predicting a destination area of ​​the moving object detected from the first image using past information indicating a past detection position of the moving object on the road; detecting a moving object from a second image included in the time series images and captured at a time later than the time when the first image was captured; A program for causing a computer to execute a process including tracking the moving object detected from the first image and the moving object detected from the second image as the same moving object when the moving object detected from the second image is detected in the area of ​​the predicted destination of the moving object detected from the first image.