JP2021117768A - Information processing device, information processing method, and program - Google Patents

Abstract

To provide an information processing device capable of detecting and notifying suspicious behavior such as moving to a shadow while avoiding public eyes while reducing the processing load, an information processing method, and a program.SOLUTION: An information processing device 101 includes: a detection unit 131 that acquires the position of an object in a captured image; a calculation unit 134 that acquires weighted distance between a first object and a second object considering information representing the position of an obstacle positioning between the first object and the second object; and a solo-action determination unit 135 that determines whether the first object is performing a solo-action different from the second object based on the weighted distance.SELECTED DRAWING: Figure 1

Description

Information processing equipment, information processing methods, and programs.

Conventionally, a video analysis system has been developed that analyzes an unspecified number of people reflected on a network camera (hereinafter referred to as a camera) and notifies a security guard or a clerk when a person who is suspicious is found. Has been done. In Patent Document 1, a monitoring area is set in one room of a nursery school, and the number of people entering and exiting across the boundary of the monitoring area is counted while distinguishing between an adult and a child based on the physique of the person, and the state of one child. If so, a method of reporting that it is in a state of caution is disclosed.

JP-A-2015-176198

The process of detecting and notifying suspicious behavior such as moving to the shadow while avoiding the eyes by image analysis of the camera image has a higher processing load than the person detection, and many calculations are performed when the number of people in the image increases. Resources are needed.

An object of the present invention is to reduce the processing load of the information processing apparatus.

In order to achieve the object of the present invention, for example, the information processing apparatus according to one embodiment has the following configurations. That is, the information indicating the position of the obstacle existing between the first object and the second object is taken into consideration with the first acquisition means for acquiring the position of the object in the captured image. Based on the second acquisition means for acquiring the weighted distance between the first object and the second object and the weighted distance, the first object is the second object. It is characterized by comprising a first determination means for determining whether or not a person is acting alone.

Reduce the processing load of the information processing device.

The figure which shows an example of the functional structure of the information processing apparatus which concerns on Embodiment 1. FIG. The figure which shows the functional structure of the computer which concerns on Embodiment 1. FIG. The figure for demonstrating the behavior of one person in the information processing apparatus which concerns on Embodiment 1. The figure for demonstrating the logical distance used by the information processing apparatus which concerns on Embodiment 1. The figure for demonstrating the movement path used by the information processing apparatus which concerns on Embodiment 1. The figure for demonstrating the position of the person used by the information processing apparatus which concerns on Embodiment 1. FIG. The figure for demonstrating the weight map used by the information processing apparatus which concerns on Embodiment 1. The figure for demonstrating the distance map used by the information processing apparatus which concerns on Embodiment 1. FIG. 5 is a flowchart showing an example of processing performed by the information processing method according to the first embodiment. The figure for demonstrating the difference between the information processing apparatus which concerns on Embodiment 2 and a conventional system. The figure which shows an example of the functional structure of the information processing apparatus which concerns on Embodiment 2. FIG. FIG. 5 is a flowchart showing an example of processing performed by the information processing method according to the second embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. The following embodiments do not limit the invention according to the claims. Although a plurality of features are described in the embodiment, not all of the plurality of features are essential to the invention, and the plurality of features may be arbitrarily combined. Further, in the attached drawings, the same or similar configurations are given the same reference numbers, and duplicate explanations are omitted.

[Embodiment 1]
FIG. 1 is a block diagram showing an example of the functional configuration of the information processing apparatus according to the present embodiment. The information processing device 101 according to the present embodiment acquires a weighted distance between a person and another person in the captured image in consideration of the movable area of the person, and the person acts alone based on the acquired weighted distance. To determine if you are doing. Therefore, the information processing device 101 includes an imaging unit 111, a recording unit 112, an image storage unit 113, a map generation unit 121, a recording unit 122, a detection unit 131, a route generation unit 132, a route storage unit 133, a calculation unit 134, and one person. It has an action determination unit 135 and an input / output unit 141. Each functional unit of the information processing apparatus 101 is realized by one or a plurality of servers, and each may be connected via LAN 151. In addition, the single-person behavior is assumed to be an behavior that may be a suspicious behavior (suspicious behavior) that avoids the public eye, and details such as processing in the determination will be described later.

The imaging unit 111 is composed of a plurality of cameras and captures an image or a moving image (hereinafter referred to as an image). The recording unit 112 is connected to the imaging unit 111, and transmits the video acquired from the imaging unit 111 to the video storage unit 113 via the LAN 151. The video storage unit 113 stores the video received from the recording unit 112.

The map generation unit 121 creates a distance map representing a weighted distance between people in the imaging range of each imaging unit in consideration of the movable area of the person, based on the behavior information of the person stored in the path storage unit 133, which will be described later. Generate. Therefore, the map generation unit 121 may generate a weight map representing the weights set for each partial region in the captured image based on the behavior information of the person. In that case, the map generation unit 121 can generate a distance map for each person based on the generated weight map and the coordinates of the person. Details of the weight map and the distance map will be described later with reference to FIGS. 4 to 8. The behavior information of the person is information about the behavior of each person detected in the image, and is not particularly limited. However, in the following, information on the position of the person for each unit time within a predetermined time (details will be described later in FIG. 6). ) Shall be used. The recording unit 122 stores the distance map generated by the map generation unit 121.

The detection unit 131 detects a person from each frame of the video stored in the video storage unit 113. The route generation unit 132 generates the movement route data of each detected person from the detection result by the detection unit 131, and transmits the movement route data to the route storage unit 133. In the following, the movement path data shall represent the time in each frame of the video and the XY coordinates of each detected person. The route storage unit 133 stores the movement route data received from the route generation unit 132. The calculation unit 134 refers to the information on the position of a person included in the movement route data stored in the route storage unit 133 based on the distance map generated by the map generation unit 121, and thereby detects between a plurality of people. Get the weighted distance of. The one-person behavior determination unit 135 determines whether or not the detected person is performing one-person behavior based on the weighted distance acquired by the calculation unit 134. In the following, for the sake of simplicity, such a judgment will be referred to as a one-person action judgment. The input / output unit 141 receives an output to a user (for example, a security guard) such as a notification of a person who is determined to have performed a single action, or an input from the user such as a video reproduction instruction.

FIG. 2 is a block diagram showing an example of the hardware configuration according to the present embodiment. The HDD 202 or ROM 203 stores an application program that operates on an operating system (OS) that performs operations such as viewing video data or analyzed data and performing various settings. The CPU 205 reads the OS and the application program from the HDD 202 or the ROM 203, loads them into the RAM 204, executes them, and advances various processes to realize each operation. The application program gets the user's input from the input device 207 connected to the computer. Further, the application program outputs various information to the output device 206 and displays the processing result. Further, the application program communicates with another computer, server, device or the like connected to the network via the communication device 208. These hardware components may be connected to each other by bus 201 and configured to be operable from an application program.

FIG. 3 is a diagram for explaining an example of one-person behavior reflected in a camera image installed in a retail store. In FIG. 3, four persons A to D (311 to 314) and product shelves 321 and 322 are shown in the image 301 of the camera of the retail store, and the person B312 moves in the shadow and performs suspicious behavior. .. In this figure, the person A311 facing the right side of the screen can visually recognize the person D314, but the person B312 and the person C313 cannot be visually recognized because the visibility is blocked by the product shelves 321 and 322, respectively. Similarly, the person C313 and the person D314 who turn to the left of the screen cannot see the person B312 because the view is blocked by the product shelf 323 even if they turn to the back. If the orientation of the face is not taken into consideration, whether or not the master-slave who visually recognizes between the two parties can be visually recognized does not change, so that the person B312 is not visually recognized by another person in the image. You will be in a position where you cannot see other people.

Here, we introduce the concept of logical distance, which is a weighted distance, which is used to judge one-person behavior. FIG. 4 is a diagram schematically showing the position of the person in FIG. In the following, for the sake of simplicity, the orientation of the person's face is not considered, and the person is illustrated with a circle symbol. The values representing the logical distances shown in FIG. 4 and the number of color-coded subregions of the three colors are convenient values for explanation, and are not particularly limited thereto, and detailed description will be described later. .. Further, FIGS. 4 to 8 below are diagrams showing the results of processing the same video, and objects having a common reference number represent the same object.

In the image 401, the positions of the persons A to D (411 to 414) are shown, and each of them is connected by a straight line together with the value of the logical distance between the two. Further, in FIG. 4, three colors are color-coded to indicate the possibility that an obstacle exists in the area, and the low area 421 where the possibility that the obstacle exists is low, and the possibility that the obstacle exists is medium. A medium area 422 and a high area 423 where obstacles are likely to be present are shown. Further, the low area 421, the middle area 422, and the high area 423 are each set to have a larger value as the possibility of the presence of an obstacle is higher, and these weights are set when calculating the logical distance. Used. In this example, the logical distance is calculated by integrating the length value of each area of the line segment between the two parties and the weight value of each area. According to such a process, it is possible to calculate a weighted logical distance in consideration of the existence of an obstacle. For example, since the distance between the person C413 and the person D414 is short and the straight line between the two passes only in the low area 421, the logical distance is a low value, which is 0.1 here. Further, although the distance between the person A411 and the person D414 is long, the logical distance is a relatively low value because the straight line between the two passes around the low area 421, which is 0.5 here. On the other hand, although the distance between the person B412 and the person D414 is short, the straight line between the two crosses the middle area 422 and the high area 423, so that the logical distance is higher than the above two values. It becomes 2.

FIG. 5 is an example of a schematic diagram in which the movement paths of four people in the video are superimposed and illustrated. On screen 501, person A is moving from right to left along the movement path 511. Similarly, the persons B, C, and D are moving along the movement paths 512, 513, and 514, respectively. Each person moves avoiding the product shelves 521 and 522, and these movement routes are represented by connecting the coordinates of the person for each unit time with a straight line.

FIG. 6 is an example of a diagram in which the coordinates of the presence of a person within a predetermined time are plotted. When the coordinates of the movement paths of a plurality of people reflected in the image at a predetermined time (for example, 1 hour) are plotted for each unit time (for example, 1 second or 0.1 second), the place where the person stays or the person enters. It is possible to read the places where there is no such place from the density of the plotted points. In particular, in this example, the blank areas and the positions of the obstacles are substantially the same, and the blank areas 611 and 612 indicated by the ellipses correspond to the product shelves 521 and 522, respectively.

FIG. 7 is an example of a weight map obtained from a movement path of a person and showing weights for each partial area in consideration of the movable area of the person. In this figure, an example in which the image is divided into 8 horizontal and 6 vertical meshes is shown, but the number of divisions of the image area is not particularly limited to this. For example, the image is divided into 16 horizontal and 9 vertical. You may. Further, the position in FIG. 7 is represented by (x, y), and for example, the position 711 of the person A is represented by (3,2). At this time, the weight w (x, y) at the position (x, y) is expressed by the following equation w.
w (x, y) = 1 / (1 + 9 × n (x, y) / max) Equation (w)

Here, n (x, y) indicates the number of points included in the mesh at the position (x, y). Further, max represents the largest value of n (x, y) in the case of taking the range of x = 1, 2, ..., 8, and y = 1, 2, ..., 6. The positions 711, 713, and 714 where the persons A, C, and D are located have a weight of 0.1 because they are places with many points, that is, places where the person often passes, while the position 712 of the person B. Has a weight of 0.4 because it is a low-traffic area.

Further, the map generation unit 121 may correct the weight of the generated weight map in consideration of the depth of the imaging region. That is, the depth on the image may be estimated from the size of the human head in the image, and the generated weight map may be corrected so that the weight assigned to the mesh at a position far from the image pickup apparatus becomes large. Therefore, the map generation unit 121 can generate a correction weight map in which the weight is likely to increase as the mesh captures the position far from the image pickup apparatus as the image pickup range. Next, the map generation unit 121 may correct the weight map by adding the weights of the generated correction weight map to the weight map generated as shown in FIG. 7 for each mesh.

FIG. 8 is an example of a distance map showing the logical distance to each position starting from the position 811 (3, 2) of the person A. Here, the logical distance from the position (x1, y1) to the position (x2, y2) is represented by d [(x1, y1), (x2, y2)]. That is, the distance map 801 shown in FIG. 8 is represented by d [(3,2), (x2, y2)] (x2 = 1, 2, ..., 8, y2 = 1, 2, ..., 6). It is a two-dimensional array. For example, the logical distance from the position 811 of the person A to the position 812 of the person B is d [(3,2), (8,4)], which is 2.1 here. Similarly, the logical distance of the person C to the position 813 is d [(3,2), (4,5)], which is 2.2 here, and the logical distance of the person D to the position 814 is d [(. 3,2), (5,5)], which is 0.5 here. Similarly, the logical distance of the person B from the position 812 is d [(8,4), (x2, y2)] (x2 = 1,2, ..., 8, y2 = 1,2, ..., 6). It is a two-dimensional array represented. In this way, a four-dimensional array can be calculated in advance from all combinations of the start point and the end point, and by referring to the distance map d [(x1, y1), (x2, y2)], the distance between the two positions can be immediately calculated. You can get the logical distance of. The size of the array of this distance map is at most 2304 (= 8 * 6 * 8 * 6), and the one-person action determination unit 135 makes a one-person action determination while easily referring to this distance map on the information processing device 101. It can be carried out.

FIG. 9 is a flowchart showing an example of processing performed by the information processing apparatus 101. In step S901, the map generation unit 121 acquires the movement route data of all the persons within the predetermined time stored in the route storage unit 133. In step S902, the map generation unit 121 collates the mesh obtained by dividing the captured image at equal intervals with the movement path data, and counts the number of times a person has existed for each mesh. In step S903, the map generation unit 121 generates a weight map in which the number of times a person has existed for each mesh is normalized between 0.1 and 1.0. In step S904, the map generation unit 121 calculates the distances between all the meshes from the weight map, and then generates a distance map.

According to such a configuration, it is possible to acquire the logical distance between the persons in consideration of the movable area of the person. By using the logical distance acquired in this way, it is possible to determine the behavior of one person regarding the person. Therefore, although the details will be described in the second embodiment described later, the system load in the information processing apparatus is reduced by reducing the number of objects for which the suspicious behavior determination, which is a high-addition processing, is performed by the one-person behavior determination, which is a low-load processing. be able to.

[Embodiment 2]
FIG. 11 is a block diagram showing an example of the functional configuration of the information processing apparatus according to the second embodiment. The information processing device 1101 according to the present embodiment determines each person in the video to act alone using the distance map generated in the same manner as in the first embodiment, and determines that the person is acting alone. Judge whether or not a person is performing suspicious behavior. That is, a person who performs a single action is extracted by a low-load image analysis process (single action determination) performed in advance, and a high-load image analysis process (suspicious behavior determination) is performed only on the extracted person. The information processing device 1101 has the same configuration as the information processing device 101 of the first embodiment except that it has the suspicious behavior determination unit 1136, and further has the same hardware configuration as shown in FIG. 2, so that the information processing device 1101 is duplicated. The description to be made is omitted.

The suspicious behavior determination unit 1136 determines whether or not the suspicious behavior is being performed for all the persons determined by the one-person behavior determination unit 135 to be performing suspicious behavior from the video stored in the video storage unit 113. Make a judgment. Here, the method of determining whether or not each person is performing suspicious behavior is not particularly limited. The suspicious behavior determination unit 1136 determines that a person who satisfies the conditions such as hiding in the shadow, putting his / her hand in his / her pocket, looking around, and being collated with the blacklist is a person who has performed suspicious behavior. You may. For the sake of simplicity, such a judgment will be referred to as a suspicious behavior judgment in the following. The input / output unit 141 may output to a user (for example, a security guard) such as threatening a person determined to have performed a suspicious action or notifying a person determined to have performed a suspicious action, and reproduces a video. Input from the user such as an instruction may be accepted. Here, the threat to a person is, for example, to play a voice such as "Please call the clerk if you have a product you are looking for" to a customer who is trying to shoplift, and use an information processing device. It refers to notifying that it is being monitored from beginning to end.

FIG. 12 is a flowchart showing an example of processing performed by the information processing apparatus 1101 according to the present embodiment. In step S1201, the detection unit 131 detects all persons by a known method from the image of the real-time image at a specific time. In step S1202, the one-person action determination unit 135 activates a subroutine described later that extracts all the persons who are performing one-person action. In step S1203 following the processing of the subroutine, the suspicious behavior determination unit 1136 selects one person who has not yet performed the suspicious behavior determination from the persons who are performing the one-person action extracted in step S1202. In step S1204, the suspicious behavior determination unit 1136 processes the suspicious behavior determination for the person selected in step S1203. In step S1205, the suspicious behavior determination unit 1136 proceeds to step S1206 if it is determined that the person has performed suspicious behavior, and proceeds to step S1207 if it is determined that the person has not performed suspicious behavior. Proceed. In step S1206, the suspicious behavior determination unit 1136 transmits an instruction such as a threat to the suspicious person or a notification to the user (for example, a clerk) to the input / output unit 141. In step S1207, the suspicious behavior determination unit 1136 determines whether or not the suspicious behavior determination has been made for all the persons who are performing the one-person action extracted in step S1202, and if it has been performed for all, the process ends. If not, the process returns to step S1203.

In step S1211 of the subroutine called from step S1202, the one-person action determination unit 135 selects one person who has not yet performed the one-person action determination from the persons detected in the image. In step S1212, the one-person action determination unit 135 acquires the logical distance between the person selected in step S1211 and all other persons in the image by referring to the positions corresponding to each person on the distance map. In step S1213, the one-person action determination unit 135 determines whether or not all the logical distances acquired in S1212 are equal to or greater than a predetermined threshold value, and if all are equal to or greater than the threshold value, the process proceeds to step S1214, and all are equal to or greater than the threshold value. If not, the process proceeds to step S1215. Here, the value set as the threshold value is not particularly limited, and a desired value is appropriately used. In step S1214, the one-person action determination unit 135 adds a flag to the selected person that he / she has performed one-person action, and proceeds to step S1215. In step S1215, the one-person action determination unit 135 determines whether or not the one-person action determination has been performed for all the persons detected in the image. If the one-person action determination is performed for all, the process returns to the caller of the subroutine and proceeds to step S1203, otherwise the process returns to step S1211.

[Difference in processing load from the conventional system]
Hereinafter, the difference in processing load in the video analysis of suspicious behavior determination between the conventional system and the present invention will be described with reference to FIG. FIG. 10 is an example of a diagram showing a customer who is reflected in an image of a camera installed in a retail store. In FIG. 10, (A) represents a case where the number of customers is small, (B) and (C) represent a case where the number of customers is large, and (A) and (B) are examples of video analysis by the conventional system. (C) shows an example of the image analysis according to the present disclosure. For the person detected, the video analysis for determining the suspicious behavior is performed by the conventional system in FIGS. 10A and 10B, and by the suspicious behavior determination unit 1136 in FIG. 10C. Although the processing load of person detection is low, the processing load of video analysis of suspicious behavior judgment is often high, and the number of targets for video analysis of suspicious behavior judgment that can be executed simultaneously by the conventional system may be limited. The explanation will be given assuming that the upper limit is 5 people.

In FIG. 10A, four people are shown in the image 1001, and a detection frame 1003 is given to the person 1002. Here, detection frames are assigned to all four persons, and video analysis of suspicious behavior determination is performed on all of them. On the other hand, in FIG. 10B, 10 people are shown in the image 1021, but there is a person such as the person 1022 who is not given a detection frame. This is because the number of detection frames granted has reached the upper limit of 5 people for video analysis, so that the person detection has been terminated at that time. In this way, if an attempt is made to perform processing with a high processing load on all objects, there is a possibility that a suspicious person who should originally perform video analysis may be overlooked.

Next, an example of the present disclosure in which the suspicious behavior determination is performed only for a person who is acting alone will be described. In FIG. 10C, the image 1041 shows 10 people as in the image 1021. In this case, the information processing device 101 assigns a detection frame to two persons, the person 1042 and the person 1043, by determining whether or not the person is acting alone, and then performs a video analysis of the suspicious action determination on these two persons. .. According to such processing, it is possible to narrow down the persons by the determination of the one-person behavior having a low processing load and reduce the number of times of video analysis of the suspicious behavior determination having a high processing load.

Further, the map generation unit 121 according to the first embodiment generates a weight map based on the movement path (position for each unit time) of the person within a predetermined time, but the weight map generation means is particularly suitable for this. Is not limited. The map generation unit 121 may acquire the position of the obstacle and generate a weight map based on the acquired position of the obstacle. The method of acquiring the position of the obstacle is not particularly limited, and for example, the user may manually input it, it may be measured by a three-dimensional sensor, it may be detected by object detection by artificial intelligence, and it may be detected in the store. It may be obtained by reading the arranged two-dimensional bar code.

Further, the map generation unit 121 generates a distance map before the operation of the system for determining suspicious behavior, but the timing at which the distance map is generated is not particularly limited to this, and the distance map is generated based on a predetermined condition. You may update it. For example, the map generation unit 121 may update the distance map at predetermined time intervals, and updates the distance map when the position, orientation, or setting of the imaging device that acquires the image to be analyzed is changed. It may be updated according to the change of the background area excluding the person in the image. That is, it may be updated at regular intervals during system operation, when the angle of view of the camera changes during system operation, or when the layout in the store changes during system operation. .. Further, the map generation unit 121 may periodically update the distance map when the processing load is low, such as at night.

Further, the number of distance maps stored in the recording unit 122 is not particularly limited. When the recording unit 122 stores a plurality of generated distance maps, the one-person action determination unit 135 determines the time, day of the week, or season when the image to be detected is captured, or depending on the situation such as an event being held. , The distance map to be referred to may be used properly.

Further, although the one-person action determination unit 135 has been described as performing the processing of steps S121 to S1215 described above for one image in the video, the number of frames for performing the one-person action determination is not particularly limited to this. For example, the one-person action determination unit 135 performs the logical distance acquisition process performed in step S1212 with respect to the images in a series of several frames before and after the image used in step S1201, and the shortest or longest logical distance in the several frames. May be used to perform each subsequent process.

Further, in step S1201, the detection unit 131 detects a person from the image, and in the subsequent processing, video analysis is performed on the person detected here, but the person to be processed is particularly limited to this. Do not mean. For example, the detection unit 131 may estimate the position of a person who has moved out of the angle of view or the position of a person hidden behind an obstacle based on the movement path in the front and rear frames. In such a case, the subsequent processing for the person detected in step S1201 is also performed for the position estimated here. According to such a process, it is possible to easily determine whether or not a person is acting alone even for a person who has moved out of the range of the image or for a person whose data is lost due to a detection failure.

(Other Examples)
The present invention supplies a program that realizes one or more functions of the above-described embodiment to a system or device via a network or storage medium, and one or more processors in the computer of the system or device reads and executes the program. It can also be realized by the processing to be performed. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

The invention is not limited to the above embodiments, and various modifications and modifications can be made without departing from the spirit and scope of the invention. Therefore, a claim is attached to make the scope of the invention public.

101: Information processing device, 111: Imaging unit, 112: Recording unit, 113: Video storage unit, 121: Map generation unit, 122: Recording unit, 131: Detection unit, 132: Route generation unit, 133: Route storage unit, 134: Calculation unit, 135: One-person behavior judgment unit, 141: Input / output unit, 151: LAN

Claims (20)

The first acquisition means for acquiring the position of the object in the captured image, and
Between the first object and the second object in consideration of the information indicating the position of the obstacle existing between the first object and the second object included in the object. A second acquisition method for acquiring the weighted distance,
Based on the weighted distance, a first determining means for determining whether or not the first object is performing a solo action different from the second object, and
An information processing device characterized by being equipped with. The weighted distance between the first object and the second object is such that the obstacle exists when there is an obstacle between the first object and the second object. The information processing apparatus according to claim 1, wherein the information processing device is defined to be larger than the case where the information processing device is not used. Further, a third acquisition means for acquiring a weight map representing the weight set for each partial region in the captured image is provided.
The second acquisition means is based on the length of the line segment between the first object and the second object in each of the partial regions and the weight set in each partial region. The information processing apparatus according to claim 1 or 2, wherein the weighted distance is acquired. The weight for each of the partial regions in the captured image is set according to whether or not the obstacle is reflected in the partial region or the possibility that the obstacle is reflected in the partial region. Item 3. The information processing apparatus according to item 3. The information processing apparatus according to claim 3 or 4, wherein the third acquisition means generates the weight map based on the movement path of the object. Any one of claims 3 to 5, wherein the third acquisition means acquires the weight map based on the position where the object exists at each of a plurality of times within a predetermined time. The information processing device described in. The information processing apparatus according to any one of claims 3 to 6, wherein the third acquisition means generates the weight map according to a user input. The information according to any one of claims 3 to 7, wherein the weight for each partial region in the captured image is further set according to the size of the object reflected in the partial region. Processing equipment. Further comprising a generation means for generating a distance map representing a weighted distance from one subregion to another in the captured image.
The second acquisition means obtains the weighted distance corresponding to the first partial region including the first object and the second partial region including the second object from the distance map. The information processing apparatus according to any one of claims 1 to 8, wherein the information processing apparatus is to be acquired. The generation means generates a plurality of the distance maps,
The second acquisition means obtains the weighted distance corresponding to the first partial region and the second partial region in the distance map selected according to the situation from the plurality of generated distance maps. The information processing apparatus according to claim 9, wherein the information processing apparatus is to be acquired. The information processing apparatus according to claim 9 or 10, wherein the generation means updates the distance map based on a predetermined condition. The information processing apparatus according to claim 11, wherein the generation means updates the distance map at predetermined time intervals. The information processing device according to claim 11 or 12, wherein the generation means updates the distance map in response to a change in the position, posture, or setting of the image pickup device that acquires the captured image. The information according to any one of claims 11 to 13, wherein the generation means updates the distance map in response to a change in the background region excluding the object in the captured image. Processing equipment. Further provided with an inferring means for estimating the position of the object which is not detected in the captured image,
The information processing apparatus according to any one of claims 1 to 14, wherein the first acquisition means acquires the position of the estimated object as the position of the object. Claims 1 to 15, further comprising a second determination means for determining whether or not the first object determined to be acting alone is a suspicious person. The information processing device according to any one item. The first acquisition means for acquiring the position of the object in the captured image, and
Between the first object and the second object in consideration of the information indicating the position of the obstacle existing between the first object and the second object included in the object. A second acquisition method for acquiring the weighted distance,
Based on the weighted distance, a first determining means for determining whether or not the first object is to be subjected to image analysis processing, and
An information processing device characterized by being equipped with. The process of acquiring the position of the object in the captured image and
Between the first object and the second object in consideration of the information indicating the position of the obstacle existing between the first object and the second object included in the object. The process of getting the weighted distance and
A step of determining whether or not the first object is acting alone differently from the second object based on the weighted distance.
An information processing method characterized by being provided with. The process of acquiring the position of the object in the captured image and
Between the first object and the second object in consideration of the information indicating the position of the obstacle existing between the first object and the second object included in the object. The process of getting the weighted distance and
A step of determining whether or not the first object is subject to image analysis processing based on the weighted distance, and
An information processing method characterized by being provided with. A program for causing a computer to function as each means of the information processing apparatus according to any one of claims 1 to 17.

Images