JP7031991B2 - Position selection device, position selection method, and position selection program - Google Patents

Description

The present invention relates to a position selection device for selecting the position of an object, a position selection method, and a position selection program.

Conventionally, for the purpose of marketing, monitoring, etc., it has been performed to track an object such as a person who stays in a predetermined area such as in a store. When there is an obstacle such as a wall or furniture that hinders the movement of a person in a predetermined area, it is expected to generate a movement flow line that does not move on the obstacle. In addition, if the position of the obstacle is obtained in advance, it is possible to track the object with high accuracy.

For example, in Patent Document 1, when tracking an object using a particle filter, if the predicted position of each particle is on the obstacle area, the predicted position of the object is corrected to a position other than the obstacle area. The method of doing so is disclosed. Further, Patent Document 2 discloses a method of lowering the existence probability of each particle on an obstacle by giving a penalty to the predicted position of each particle according to the height of the obstacle.

Japanese Patent No. 5687035 Japanese Patent No. 5552069

Conventionally, in order to realize object tracking, it has been the mainstream to use an approximate filter such as a Kalman filter or a particle filter. In recent years, with the development of machine learning and deep learning technologies (for example, Support Vector Machine and Convolutional Neural Network), the Tracking-by-Detection method has attracted attention as a method for tracking objects with high accuracy from images captured in a predetermined area. I'm bathing. In the Tracking-by-Detection method, the position of an object in an image captured at each time is detected using an object detector (Detection), and the object detected at each time is detected using the same object detector (Re-Identification). By associating with, tracking of an object is realized.

FIG. 15 is a diagram showing an example of tracking an object by the Tracking-by-Detection method in a predetermined area where an obstacle area exists. Here, it is assumed that the position of the obstacle area is known in advance.

The tracking result of the object up to the time t-1 is the detection result of the object at the time before the time t-1, for example, the time t-4, the time t-3, and the time t-2, and the detection of the object at the time t-1. Generated by associating with the result. When tracking the object up to time t, the problem of associating the tracking result up to time t-1 with the detection result at time t, that is, the detection result A or the detection result B is solved.

At this time, an obstacle region exists on the line segment connecting the object position at time t-1 included in the tracking result and the object positions of the detection result A and the detection result B at time t. However, due to the positional relationship between the position of the object and the obstacle area, it is highly possible that the object is detected as the detection result A at time t, and the tracking result of the object is set to the detection result A rather than the detection result B. It is desirable to associate them preferentially. In other words, in the Tracking-by-Detection method, which simply associates the tracking result with the detection result in chronological order, the correspondence between the tracking result and the detection result is not taken into consideration because the obstacle area existing in the middle is not considered. There is a problem that it cannot be identified correctly.

On the other hand, the A * (Aster) algorithm is known as a method for searching the shortest path. In FIG. 15, by applying the A * algorithm with the object position at time t-1 included in the tracking result as the start point and the object position at time t (each of the detection results A and B) as the end point, the start point It is possible to identify a detour (shortest route) that avoids obstacles and reaches the end point. Therefore, by using the length on the route of the detour as a cost value for specifying the correspondence between the object position at time t-1 and the object positions of the detection result A and the detection result B at time t, It is possible to preferentially associate the tracking result of the object with the detection result A over the detection result B.

However, the A * algorithm has a problem that the amount of calculation is large. In particular, in the Tracking-by-Detection method, when tracking multiple objects at the same time, (the number of objects being tracked by time t-1) x (the number of objects detected at time t) when identifying the correspondence. It is necessary to apply the A * algorithm by the amount of), and there is a problem that the amount of calculation becomes enormous.

Therefore, the present invention has been made in view of these points, and an object of the present invention is to provide a position selection device, a position selection method, and a position selection program capable of efficiently specifying the correspondence between objects. ..

The position selection device according to the first aspect of the present invention is an obstacle area specifying unit that specifies an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area. An object position specifying unit that specifies a position of an object moving in the predetermined area before the first time, and a candidate for one or more detection positions of the object at the second time, which is a time after the first time. The positional relationship between the detection position specifying unit that specifies the candidate position, the specified position of the object before the first time, the specified one or more candidate positions, and the specified position of the obstacle area. Based on this, a movement obstruction degree specifying unit that specifies a movement obstruction degree indicating difficulty when the object moves to each of the one or more candidate positions, and one or more from the position of the object before the first time. A selection unit that specifies the movement cost to each of the candidate positions based on the movement obstruction degree and selects the candidate position having a relatively small movement cost as the position of the object at the second time. Be prepared.

The object position specifying unit predicts the position of the first object, which is the position of the object at the second time, based on the position of the object before the first time, and the movement obstruction degree specifying unit predicts the position of the first object. The movement obstruction degree may be specified based on the positional relationship between one object position, one or more specified candidate positions, and the specified position of the obstacle area.

The object position specifying unit specifies the position of the object at the first time as the first object position, and the movement obstruction degree specifying unit specifies the first object position and one or more specified candidate positions. The degree of movement obstruction may be specified based on the positional relationship with the position of the obstacle region.

The movement obstruction degree specifying unit specifies a line segment connecting the first object position and each of the specified one or more candidate positions, and is based on the positional relationship between the line segment and the specified obstacle area. Then, the movement obstruction degree corresponding to each of the specified one or more candidate positions may be specified.

The movement obstruction degree specifying unit identifies two intersections of the line segment and the specified obstacle region for each of the specified one or more line segments, and the specified two intersections and the obstacle area. The degree of movement obstruction may be specified based on the relationship with.

Even if the movement obstruction degree specifying unit specifies the movement obstruction degree based on the length of a relatively short contour line among the two contour lines of the obstacle region divided by the two specified intersections. good.
The movement obstruction degree specifying unit can be taken from among three intersections for each of the specified one or more line segments when there are three or more intersections between the line segment and the specified obstacle region. Of the two intersections, two intersections having a relatively long distance may be specified.

The movement obstruction degree specifying unit is based on the ratio of the length of the contour line of the obstacle region divided by the two specified intersections to the length of the line segment connecting the two specified intersections. May be specified.
The movement obstruction degree specifying unit specifies the angle formed by each of the two specified intersections and the two line segments connecting the specified obstacle region reference point, and determines the movement obstruction degree based on the angle. It may be specified.

The movement obstruction degree specifying unit identifies two sides in which any of the two specified intersections exists among the sides constituting the specified obstacle area, and two side groups existing between the two sides. Of these, the degree of movement interference may be specified based on the number of sides included in the side group having a relatively small number of sides.
When there are a plurality of the obstacle regions having intersections with the specified line segments, the movement obstruction degree specifying unit may approximate the plurality of obstacle regions to one obstacle region.

When the first object position exists in the obstacle area, the object position specifying unit may modify the first object position to a position outside the obstacle area.
When the shape of the obstacle is non-convex, the obstacle region specifying portion may approximate the obstacle region corresponding to the obstacle to the convex obstacle region.
When a plurality of the specified obstacle areas are specified and the specified plurality of obstacle areas are adjacent to each other or overlapped with each other, the obstacle area specifying unit can use the plurality of obstacle areas as the plurality of obstacle areas. It may be approximated to the sum region of.
When the specified candidate position exists in the obstacle region, the detection position specifying unit may modify the candidate position to a position outside the obstacle region.

The selection unit specifies the degree of similarity in a predetermined correspondence between the specified object corresponding to the position before the first time and the object corresponding to each of the specified one or more candidate positions, and the degree of similarity. And, based on the specified movement obstruction degree, the movement cost from the position before the first time to each of the one or more candidate positions may be specified.

The selection unit may control not to select the candidate position when the relatively small movement cost exceeds the first threshold value.
The selection unit may set the movement cost to a value exceeding the first threshold value for the candidate position where the specified movement interference degree exceeds the second threshold value.

The position selection device further provides a flow line generation unit that generates a movement line of the object by associating the position of the object at the first time with the candidate position of the object selected by the selection unit. You may prepare.

The position selection method according to the second aspect of the present invention specifies an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area, which is executed by a computer. A step, a step of specifying a position of an object moving in a predetermined area before the first time, and a candidate for one or more detection positions of the object at a second time, which is a time after the first time. The object is based on the positional relationship between the step of specifying the candidate position, the position of the object before the specified first time, the specified one or more candidate positions, and the position of the specified obstacle area. The step of specifying the degree of movement obstruction indicating the difficulty when moving to each of the one or more candidate positions, and the cost of moving the object from the position before the first time to each of the one or more candidate positions. Is specified based on the degree of movement obstruction, and includes a step of selecting the candidate position where the movement cost is relatively small as the position of the object at the second time.

The position selection program according to the third aspect of the present invention causes a computer to identify an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area. Area specifying unit, object position specifying unit that specifies the position of the object moving in the predetermined area before the first time, and one or more detection positions of the object at the second time, which is a time after the first time. The positional relationship between the detection position specifying unit that specifies the candidate position that is a candidate, the position of the object that was specified before the first time, the specified one or more candidate positions, and the position of the specified obstacle area. Based on this, the movement obstruction degree specifying unit that specifies the movement obstruction degree indicating the difficulty when the object moves to each of the one or more candidate positions, and one or more from the position of the object before the first time. As a selection unit, which specifies the movement cost to each of the candidate positions based on the movement obstruction degree, and selects the candidate position where the movement cost is relatively small as the position of the object at the second time. Make it work.

According to the present invention, there is an effect that the correspondence between objects can be efficiently specified.

It is a figure explaining the outline of the flow line generation apparatus which concerns on this embodiment. It is a figure which shows the structure of the flow line generation apparatus which concerns on this embodiment. It is a figure which shows the example which specified the position and the candidate position of the object at the 1st time. It is a figure which shows the example which specified the line segment which connects the 1st object position and each of one or more candidate positions. It is a figure which shows the example which specified the intersection of the line segment connecting the 1st object position and the candidate position, and the obstacle area specified as a rectangle. It is a figure which shows the example of two contour lines of the obstacle area divided by the intersection of two specified points. It is a figure which shows the example of calculating the ratio of the length of the contour line to the length of the line segment connecting two intersections. It is a figure which shows the example which specifies the angle formed by each of the two specified intersections, and the two line segments connecting with the reference point of the specified obstacle area. It is a figure which shows the example which specifies the movement obstruction degree based on the number of sides. It is a figure which shows the example which approximates a plurality of obstacle areas to one obstacle area. It is a figure which shows the example which specified the correspondence relation between the position of the object before the 1st time, and the candidate position. It is a figure which shows the example which calculates the movement obstruction degree based on the ratio of the length of the contour line to the length of the line segment connecting two intersections. It is a figure which shows the example which specified the fixed position corresponding to the 2nd time. It is a flowchart which shows the flow of the process in the flow line generation apparatus which concerns on this embodiment. It is a figure which shows the example of the case where the object is tracked by the Tracking-by-Detection method in the predetermined area where the obstacle area exists.

[Overview of flow line generator 1]
FIG. 1 is a diagram illustrating an outline of a flow line generation device 1 according to the present embodiment. The flow line generation device 1 is a computer that generates flow line information indicating the position of an object moving in a predetermined area at each time, and functions as a position selection device that selects the position of the object from the candidate positions of the object.

Here, the predetermined area is, for example, the floor of a store. In the predetermined area, a person such as a clerk exists as a predetermined object to be analyzed for movement. In the following description, a predetermined object is simply referred to as an object.

The flow line generation device 1 specifies an obstacle region corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle region in the predetermined area. The flow line generator 1 identifies a position of an object moving in a predetermined area before the first time, and is a candidate for one or more detection positions of the object at the second time, which is a time after the first time. Identify the location. FIG. 1 shows an example in which the flow line generator 1 identifies two candidate positions P2A and P2B.

In the flow line generator 1, the object is set before the first time based on the positional relationship between the specified position of the object before the first time, the specified one or more candidate positions, and the position of the specified obstacle area. Identify the degree of obstruction that indicates the difficulty of moving from one position to each of one or more candidate positions. In the example shown in FIG. 1, the flow line generation device 1 specifies that the movement obstruction degree of the candidate position P2A is lower than the movement obstruction degree of the candidate position P2B.

Then, the flow line generation device 1 specifies the movement cost from the position at the first time to each of the one or more candidate positions based on the degree of movement obstruction, and the candidate position where the movement cost is relatively small is the first object. Select the position at 2 time. In the example shown in FIG. 1, the flow line generator 1 specifies the movement cost to the candidate position P2A to be lower than the movement cost to the candidate position P2B, and selects the candidate position P2A as the position of the object at the second time. do. By doing so, the flow line generator 1 is more efficient than the case where the shortest path from the position at the first time to each of the plurality of candidate positions is specified by using a tracking algorithm such as A *. Candidate positions can be selected.

[Structure of flow line generator 1]
Subsequently, the configuration of the flow line generation device 1 will be described. FIG. 2 is a diagram showing a configuration of a flow line generation device 1 according to the present embodiment. As shown in FIG. 2, the flow line generation device 1 includes a storage unit 11 and a control unit 12.

The storage unit 11 is a storage medium including a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The storage unit 11 stores a program executed by the control unit 12. For example, the storage unit 11 uses the flow line generation device 1 as an obstacle area identification unit 121, an object position identification unit 122, a detection position identification unit 123, a movement obstruction degree identification unit 124, a selection unit 125, and a flow line generation unit 126. It stores a flow line generation program as a position selection program that functions as a position selection program.

The control unit 12 is, for example, a CPU (Central Processing Unit). By executing the flow line generation program stored in the storage unit 11, the control unit 12 executes the obstacle area identification unit 121, the object position identification unit 122, the detection position identification unit 123, the movement obstruction degree identification unit 124, and the selection unit. It functions as 125 and a flow line generator 126.

The obstacle area specifying unit 121 identifies an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area. For example, the storage unit 11 stores one or more obstacle area information indicating the shape of the obstacle area and the position of the obstacle area in a predetermined area. The obstacle area information is, for example, information indicating the position (global position) of each vertex of the polygon indicating the shape of the obstacle area in a predetermined area. When the polygonal area is rectangular, the obstacle area information may be position coordinates indicating the center position of the rectangular area and information indicating the horizontal width and the vertical width. The obstacle area specifying unit 121 identifies the obstacle area and the position of the obstacle area in a predetermined area by referring to the obstacle area information.

When the shape of the specified obstacle is non-convex, the obstacle region specifying unit 121 may approximate the obstacle region corresponding to the obstacle to the convex obstacle region. In addition, when a plurality of obstacle areas are specified, the obstacle area identification unit 121 can use the plurality of obstacle areas as the plurality of obstacles when the specified plurality of obstacle areas are adjacent to each other or overlap with each other. It may be approximated to the sum region of the regions. Here, the obstacle area specifying unit 121 may also specify the wall as an obstacle. When the wall is in contact with other obstacles, the obstacle area specifying unit 121 may approximate these obstacle areas to one obstacle area which is the sum area of these obstacle areas. good. By doing so, the flow line generation device 1 can simplify the shape of the obstacle region and reduce the processing load.

The object position specifying unit 122 specifies the position of the object moving in the predetermined area before the first time. For example, the storage unit 11 stores one or more flow line information indicating the position of an object moving in a predetermined area at each time. The flow line information is time-series information in which the time is associated with the position information indicating the position of the object in a predetermined area at the time. The position information is, for example, two-dimensional coordinates in units of meters or centimeters.

The position information included in the flow line information may be information capable of specifying a global position indicating the position of a predetermined area. For example, local position information indicating a position in a captured image obtained by capturing a predetermined area. And information for conversion for converting the position to a global position may be included. The local position information is, for example, rectangular information composed of two-dimensional coordinates indicating the position of an object in pixels and the width and height of the object.

The object position specifying unit 122 identifies one or more flow line information including the position information up to the first time, and by specifying the position information associated with the first time in the flow line information, one or more. Specify the position of the object at the first time. The object position specifying unit 122 may specify the flow line information that does not include the first time and includes only the position information corresponding to the time before the first time. Then, the object position specifying unit 122 may specify the position of the object before the first time by specifying the position information associated with the latest time in the specified flow line information.

Further, the object position specifying unit 122 may specify the position of the object at the first time based on the captured image captured at the first time by the image pickup device that images the predetermined area. In this case, the object position specifying unit 122 inputs the captured image captured at the first time to the object detection device that detects the position of the object based on the captured image. Then, the object position specifying unit 122 acquires position information indicating the position of one or more objects reflected in the captured image captured at the first time from the object detection device, and first obtains the position indicated by the acquired position information. Specify as the position of the object at the time. Here, the object detection device may be one using machine learning or deep learning. Further, the object detection device may be one using a 3D sensor or the like using infrared rays.

The detection position specifying unit 123 identifies a candidate position that is a candidate for one or more detection positions of the object at the second time, which is a time after the first time. For example, the detection position specifying unit 123 inputs the captured image captured at the second time to the object detection device. Then, the detection position specifying unit 123 identifies one or more candidate positions by acquiring position information indicating the positions of one or more objects reflected in the captured image captured at the second time from the object detection device.

When the specified candidate position exists in the obstacle area, the detection position specifying unit 123 may correct the candidate position to a position outside the obstacle area. For example, when the specified candidate position exists in the obstacle area, the detection position specifying unit 123 selects the position closest to the candidate position among the plurality of positions showing the outline of the obstacle area as the corrected candidate. Position. The flow line generation device 1 may control the object detection device so that the object detection device does not detect the position inside the obstacle area as the position of the object.

FIG. 3 is a diagram showing an example in which the position and the candidate position of the object before the first time are specified. In the example shown in FIG. 3, the latest object position P1A corresponding to the flow line T1 and the latest object position P1B corresponding to the flow line T2 are specified, and the candidate positions P2A, P2B and P2C are specified. You can check it. The latest object position P1A corresponding to the flow line T1 is the position of the object at the first time, and the latest object position P1B corresponding to the flow line T2 is the position of the object at the time before the first time.

The movement obstruction degree specifying unit 124 has specified the position of the object before the first time specified by the object position specifying unit 122, one or more candidate positions specified by the detection position specifying unit 123, and the obstacle area specifying unit 121. Based on the positional relationship with the position of the obstacle area, the degree of movement obstruction indicating the difficulty when the object moves to each of one or more candidate positions is specified.

First, the object position specifying unit 122 predicts the position of the first object, which is the position of the object at the second time, based on the position of the object indicated by the specified flow line information before the first time. For example, the object position specifying unit 122 predicts the position of the first object by using an approximate filter such as a Kalman filter or a particle filter.

Then, the movement obstruction degree specifying unit 124 specifies the movement obstruction degree based on the positional relationship between the position of the first object, the specified one or more candidate positions, and the position of the specified obstacle area. Specifically, the movement obstruction degree specifying unit 124 specifies a line segment connecting each of the specified first object position and one or more specified candidate positions, and the line segment and the specified obstacle area are defined. Based on the positional relationship, the degree of movement obstruction when the object moves from the first object position to each of the specified one or more candidate positions is specified.

The object position specifying unit 122 determines that the first object position is the position of the object at the second time predicted based on the position before the first time, but the present invention is not limited to this. For example, the object position specifying unit 122 may specify the position of the object at the first time as the first object position. Then, the movement obstruction degree specifying unit 124 has a movement obstruction degree based on the positional relationship between the first object position specified by the object position specifying unit 122, one or more specified candidate positions, and the position of the specified obstacle area. May be specified. By doing so, the flow line generation device 1 can specify the position of the object at the second time without predicting the position of the object at the second time.

Further, when the specified first object position exists in the obstacle area, the object position specifying unit 122 may correct the first object position to a position outside the obstacle area. By doing so, the flow line generation device 1 can prevent the position of the first object from being present in the obstacle area, thereby reducing the accuracy of specifying the degree of movement obstruction.

FIG. 4 is a diagram showing an example in which a line segment connecting each of the first object position and one or more candidate positions is specified. In FIG. 4, first object positions PPA and PPB corresponding to the second time are specified for each of the flow line T1 and the flow line T2, and one or more candidate positions P2A and P2B are specified from these first object positions. And it can be confirmed that 6 line segments for P2C have been specified.

Hereinafter, a plurality of methods for specifying the movement obstruction degree by the movement obstruction degree specifying unit 124 will be described. The movement obstruction degree specifying unit 124 specifies the movement obstruction degree when moving from each of one or more first object positions to each of one or more candidate positions based on at least one of one or more identification methods. .. For example, the movement obstruction degree specifying unit 124 specifies one movement obstruction degree corresponding to when moving from one first object position to one candidate position by combining at least one of a plurality of identification methods. May be good. When moving from one first object position to one candidate position, the movement obstruction degree specifying unit 124 specifies a plurality of movement obstruction degrees according to each of the plurality of identification methods. One movement obstruction degree may be specified by using the product of a plurality of movement obstruction degrees specified by each processing method, a weighted sum, or the like.

The movement obstruction degree specifying unit 124 identifies two intersections of the line segment and the specified obstacle area for each of the specified one or more line segments. FIG. 5 is a diagram showing an example in which the intersection of the line segment connecting the first object position and the candidate position and the obstacle area specified as a rectangle is specified. As shown in FIG. 5, the intersections i1A and i1B are the intersections of the two line segments connecting the first object position PPA and the candidate positions P2A and P2B and the sides (contours) of the obstacle region specified as a rectangle. It can be confirmed that i2A and i2B have been specified. When the obstacle region has a convex shape, there are two intersections between one line segment passing through the first object position and the candidate position and the contour of the obstacle region. When the line segment intersects the apex of the obstacle region, the intersection becomes one point. In this case, the movement obstruction degree specifying unit 124 specifies the intersection of the two points as the same point. May be good.

Further, when the obstacle region has a non-convex shape, there may be three or more intersections between the line segment and the obstacle region for each of the specified one or more line segments. In this case, the movement obstruction degree specifying unit 124 specifies two intersections having a relatively long distance from the two possible intersections among the three intersections as two intersections used for specifying the movement obstruction degree. By doing so, the flow line generation device 1 can accurately specify the degree of movement obstruction even if the obstacle region has a non-convex shape.

The movement obstruction degree specifying unit 124 specifies the movement obstruction degree based on the relationship between the two specified intersections and the obstacle area. For example, the movement obstruction degree specifying unit 124 specifies the movement obstruction degree based on the length of the relatively short contour line among the two contour lines of the obstacle region divided by the intersection of the two specified points.

FIG. 6 is a diagram showing an example of two contour lines of an obstacle region divided by an intersection of two specified points. It can be confirmed that FIG. 6 shows two contour lines LA and LB when the obstacle area is divided by the line segment connecting the first object position PPA and the candidate position P2A. In the example shown in FIG. 6, it can be confirmed that the contour line LA is shorter than the contour line LB. In the example shown in FIG. 6, the movement obstruction degree specifying unit 124 specifies the movement obstruction degree when the object moves from the first object position PPA to the candidate position P2A based on the length of the contour line LA. For example, the movement obstruction degree specifying unit 124 calculates the length of the contour line LA as the movement obstruction degree. Similarly, the movement obstruction degree specifying unit 124 obtains the shorter contour line when the obstacle area is divided by the line segment connecting the first object position PPA and the candidate position P2B, and the object is the first object position PPA. The degree of movement obstruction when moving from to the candidate position P2B is specified.

Further, the movement obstruction degree specifying unit 124 has a movement obstruction degree based on the ratio of the length of the contour line of the obstacle region divided by the two specified intersections to the length of the line segment connecting the two specified intersections. May be specified. For example, the movement obstruction degree specifying unit 124 specifies a relatively short contour line among the two contour lines when the obstacle area is divided by a line segment connecting the first object position and one candidate position. do. Then, the movement obstruction degree specifying unit 124 calculates, for example, the ratio of the length of the contour line to the length of the line segment connecting the two intersections corresponding to the contour line as the movement obstruction degree.

FIG. 7 is a diagram showing an example of calculating the ratio of the length of the contour line to the length of the line segment connecting the two intersections. FIG. 7 shows a relatively short contour line LA when the obstacle area is divided by a line segment connecting the first object position PPA and the candidate position P2A, and connecting the first object position PPA and the candidate position P2B. A relatively short contour line LC when the obstacle area is divided by a line segment is shown.

In the example shown in FIG. 7, the movement obstruction degree specifying unit 124 sets the ratio of the length of the contour line LA to the length of the line segment connecting the two intersections i1A and i2A corresponding to the contour line LA as the first object. It is specified as the degree of movement obstruction when moving from the position PPA to the candidate position P2A. Further, the movement obstruction degree specifying unit 124 sets the ratio of the length of the contour line LC to the length of the line segment connecting the two intersections i1B and i2B corresponding to the contour line LC, from the first object position PPA to the candidate position. It is specified as the degree of movement obstruction when moving to P2B.

Further, the movement obstruction degree specifying unit 124 specifies the angle formed by the two line segments connecting each of the two specified intersections and the reference point of the specified obstacle region, and determines the movement obstruction degree based on the angle. It may be specified. Here, the reference point of the obstacle region is, for example, the center point of the obstacle region.

FIG. 8 is a diagram showing an example of specifying the angle formed by two line segments connecting each of the two specified intersections and the reference point of the specified obstacle region. FIG. 8 shows a line segment connecting the two intersections i1A and i2A when the obstacle area is divided by the line segment connecting the first object position PPA and the candidate position P2A, and the first object position PPA and the candidate position P2B. The two intersections i1B and i2B when the obstacle area is divided by the above are shown.

In the example shown in FIG. 8, the movement obstruction degree specifying unit 124 specifies the angle θA between the two intersections i1A and i2A corresponding to the candidate positions P2A and the center point O. Then, the movement obstruction degree specifying unit 124 sets the angle θA as the movement obstruction degree when the object moves from the first object position PPA to the candidate position P2A. Further, the movement obstruction degree specifying unit 124 specifies the angle θB between the two intersections i1B and i2B corresponding to the candidate positions P2B and the center point O. Then, the movement obstruction degree specifying unit 124 sets the angle θB as the movement obstruction degree when the object moves from the first object position PPA to the candidate position P2B.

Further, the movement obstruction degree specifying unit 124 identifies two sides in which any of the two specified intersections exists among the sides constituting the specified obstacle area, and the two sides existing between the two sides. The degree of movement obstruction may be specified based on the number of sides included in the side group having a relatively small number of sides in the group.

FIG. 9 is a diagram showing an example of specifying the movement obstruction degree based on the number of sides. The obstacle area shown in FIG. 9 is octagonal and includes sides E1 to E8. FIG. 9 shows two intersections i1A and i2A when the obstacle area is divided by a line segment connecting the first object position PPA and the candidate position P2A. In the example shown in FIG. 9, the movement obstruction degree specifying unit 124 specifies the sides E3 and E6 as two sides where any of the two intersections i1A and i2A exists. Between the sides E3 and E6, there is a first side group including the sides E4 and E5 and a second side group including the sides E1, E2, E7 and E8. When the object moves from the first object position PPA to the candidate position P2A, the movement obstruction degree specifying unit 124 is based on the number of sides (two) included in the first side group having a relatively small number of sides. Identify the degree of movement obstruction. For example, the movement obstruction degree specifying unit 124 uses the number of sides included in the first side group, which has a relatively small number of sides, as the movement obstruction degree.

Further, when there are a plurality of obstacle regions having intersections with the specified line segments, the movement obstruction degree specifying unit 124 may approximate the plurality of obstacle regions to one obstacle region.

FIG. 10 is a diagram showing an example of approximating a plurality of obstacle regions to one obstacle region. In FIG. 10, obstacle regions X, Y, and Z exist as a plurality of obstacle regions, and these obstacle regions are divided by a line segment connecting the first object position PPA and the candidate position P2A. Six intersections i1A to i6A are shown. In this case, the movement obstruction degree specifying unit 124 approximates the three obstacle regions to one obstacle region SA. By doing so, the movement obstruction degree specifying unit 124 can treat these obstacle areas as one obstacle area, so that the load when specifying the movement obstruction degree can be reduced.

The selection unit 125 specifies the movement cost from the position before the first time of the object to each of one or more candidate positions based on the degree of movement obstruction, and the candidate position where the movement cost is relatively small is the first candidate position of the object. Select the position at 2 time.

The selection unit 125 has a predetermined correspondence relationship between the object corresponding to the position before the first time specified by the object position identification unit 122 and the object corresponding to each of the one or more candidate positions specified by the detection position identification unit 123. Identify the similarity cost that indicates the similarity in. Then, the selection unit 125 specifies the movement cost from the position before the first time to each of one or more candidate positions based on the similarity cost and the movement interference degree specified by the movement interference degree specifying unit 124. do.

When the position of the object at the first time can be specified, the selection unit 125 specifies the similarity cost between the object corresponding to the position and the object corresponding to each of the one or more candidate positions, and first. If the position of the object at the time cannot be specified, the object corresponding to the position of the object corresponding to the latest time and the object corresponding to each of one or more candidate positions before the first time are selected. Identify the similarity cost of.

In the present embodiment, the smaller the similarity cost, the more similar the object before the first time and the object corresponding to the candidate position are. The selection unit 125 calculates, for example, the linear distance or the shortest distance between the position of the object before the first time in the global coordinate system corresponding to the position of the predetermined area and each of the one or more candidate positions as the similarity cost. Further, the selection unit 125 calculates the overlap ratio (IoU: Intersection over Union) between the projection regions corresponding to the objects in the local coordinate system which is the coordinate system corresponding to the position on the captured image captured by the imaging unit. The value obtained by subtracting the duplication rate from 1 may be specified as the similarity cost.

Further, if there is an object image showing an object extracted from the captured image of a predetermined area captured by the imaging device, the selection unit 125 may specify the similarity cost based on the similarity of the object images. good.

For example, the object image corresponding to the position of the object at the first time, which is identified from the captured image of the predetermined area captured at the first time by the imaging device, is set as the first object image, and the predetermined area captured at the second time. The object image corresponding to each of the one or more candidate positions identified from the captured image is referred to as a second object image. In this case, the selection unit 125 inputs each of the first object image and one or more second object images to the same person determination device (not shown), and acquires a determination score from the same person determination device. Here, it is assumed that the higher the possibility that the person corresponding to the first object image and the person corresponding to the second object image are the same, the higher the value of the determination score. The selection unit 125 may specify, for example, the reciprocal of the determination score obtained from the same person determination device as the similarity cost.

Further, the selection unit 125 may specify the similarity cost based on the feature vector showing the features of the first object image and the feature vector showing the features of the second object image. For example, the selection unit 125 inputs a first object image and one or more second object images to the feature amount extraction device (not shown), and displays the feature vectors of these object images from the feature amount extraction device. Get information. Then, the selection unit 125 calculates the cosine similarity determined by the angle formed by the feature vector of the first object image and the feature vector corresponding to each of the one or more second object images. The selection unit 125 specifies a value obtained by subtracting the calculated cosine similarity from 1 for each of the one or more second object images as the similarity cost corresponding to the second object image.

Here, machine learning or deep learning can be used for the same person determination device and the feature amount extraction device. By combining at least one of the plurality of similarity cost specifying methods described above, the selection unit 125 has a similarity between the object corresponding to the position before the first time and the object corresponding to one candidate position. You may specify the cost.

For example, the selection unit 125 specifies the similarity cost for the object corresponding to the position before the first time and the object corresponding to one candidate position according to each of the plurality of similarity cost specifying methods. Then, the selection unit 125 uses the product of the similarity costs specified by each specific method, the weighted sum, and the like to create an object corresponding to the position before the first time and an object corresponding to one candidate position. Identify one corresponding similarity cost.

The selection unit 125 specifies the movement cost based on at least one of the specified similarity cost and the movement interference degree specified by the movement interference degree specifying unit 124. The selection unit 125 specifies the movement cost by, for example, calculating the product of the similarity cost and the movement obstruction degree and the weighted sum.

The selection unit 125 determines the movement cost in specifying the correspondence between the one or more objects specified by the object position identification unit 122 and the objects corresponding to each of the one or more candidate positions detected by the detection position identification unit 123. Generate a cost matrix as an element.

In the case of the example shown in FIG. 4, the selection unit 125 has the latest object positions P1A and P1B corresponding to the flow lines T1 and T2 specified by the object position identification unit 122 as each column, and the candidate positions P2A, P2B and P2C as each column. To generate a cost matrix with 2 rows and 3 columns. Each element of the matrix contains a travel cost value identified from the correspondence shown by each line segment in FIG. By applying the Hungarian method to the matrix, the selection unit 125 specifies the correspondence between the position of the object before the first time when the cost value is minimized and the candidate position.

FIG. 11 is a diagram showing an example in which the correspondence between the position of the object before the first time and the candidate position is specified. FIG. 11 shows the latest object positions P1A and candidates as the correspondence between the latest object positions P1A and P1B shown in FIG. 4 and the candidate positions P2A, P2B and P2C when the Hungarian method is applied. An example is shown in the case where two correspondences of the position P2A, the latest object position P1B, and the candidate position P2B are specified. It can be confirmed that the candidate position P2C does not have a correspondence with any of the positions of the plurality of latest objects.

In the above example, since a non-square matrix (matrix of 2 rows and 3 columns) is used as the cost matrix, a candidate position such as the candidate position P2C in which the correspondence is not specified occurs, but when a square matrix is used, all the candidate positions are generated. Correspondence is obtained about the position of the latest object and the candidate position. Even in such a case, the obtained correspondence is not always correct.

Further, since the movement obstruction degree is a value that simply specifies the movement cost based on the obstacle, for example, from the viewpoint of the movement distance, the magnitude relation of the movement obstruction degree is not always correct. On the other hand, in order to specify the optimum correspondence from the viewpoint of the travel distance, it is necessary to use the shortest path search algorithm with a large amount of calculation. Further, in specifying the movement cost to be included in the cost matrix, it is necessary to reflect the movement interference degree in the similarity cost, but when the weighted sum is used, it is often difficult to preset an appropriate weight parameter.

Therefore, the selection unit 125 may specify the correspondence relationship with high accuracy by performing the threshold value processing on the movement cost corresponding to the obtained correspondence relationship. Specifically, the selection unit 125 determines the object when the relatively small movement cost of the movement cost between the latest object position and each of the one or more candidate positions exceeds the first threshold value. It may be controlled not to select the corresponding candidate position.

Further, the selection unit 125 sets the movement cost corresponding to the candidate position to a value exceeding the first threshold value for the candidate position where the movement interference degree specified by the movement interference degree specifying unit 124 exceeds the second threshold value. You may.

FIG. 12 is a diagram showing an example of calculating the movement interference degree based on the ratio of the length of the contour line to the length of the line segment connecting the two intersections. For example, as shown in FIG. 12, when the first object position PPA and the candidate positions P2A, P2B, P2C and P2D are specified, the movement obstruction degree specifying unit 124 is the length of the line segment connecting the two intersections. It is assumed that the degree of movement obstruction is specified by the ratio of the length of the contour line to the height.

Assuming that the movement obstruction degree corresponding to the candidate position P2A is DA, the movement obstruction degree corresponding to the candidate position P2B is DB, the movement obstruction degree corresponding to the candidate position P2C is DC, and the movement obstruction degree corresponding to the candidate position P2D is DD. The magnitude relationship is DA <DB << DD <DC. According to the magnitude relationship, the candidate position P2A is preferentially associated with the first object position PPA over the candidate position P2B, and the candidate position P2D is associated with the first object position PPA preferentially over the candidate position P2C. Will be. However, it cannot be said that the magnitude relationship is always correct when making the correspondence.

The selection unit 125 sets the movement cost to a value exceeding the first threshold value for the candidate position where the movement interference degree exceeds the second threshold value. For example, the selection unit 125 uses the similarity cost as the movement cost when the movement interference degree is equal to or less than the second threshold value, and sets the value exceeding the first threshold value as the movement cost when the movement interference degree exceeds the second threshold value.

Further, the selection unit 125 sets the movement obstruction degree to 1 when the movement obstruction degree is equal to or less than the second threshold value, and sets the movement obstruction degree to a significantly large value when the movement obstruction degree exceeds the second threshold value. good. Then, the selection unit 125 may use the product of the movement obstruction degree and the similarity cost as the movement cost. By doing so, when the movement obstruction degree of the candidate position with respect to the first object position exceeds the second threshold value, the candidate position is controlled so as not to be selected as the candidate position corresponding to the first object. Can be done.

In the example shown in FIG. 12, the selection unit 125 sets the second threshold value to a value larger than the movement interference degree DB and smaller than the movement interference degree DD. By doing so, in the example shown in FIG. 12, the selection unit 125 prevents the candidate position P2C and the candidate position P2D from being selected with respect to the first object position PPA, and the selection unit 125 with respect to the first object position PPA. Of the candidate positions P2A and P2B, the candidate positions having the smaller similarity cost can be associated with each other.

The flow line generation unit 126 generates a movement flow line of the object by associating the position of the object at the first time with the candidate position of the object selected by the selection unit 125. For example, the flow line generation unit 126 adds the flow line to the flow line information specified by the object position specifying unit 122 in association with the second time and the position information indicating the candidate position selected by the selection unit 125. Update the information.

The flow line generation unit 126 may use the candidate position selected by the selection unit 125 as an observation value of the approximate filter. That is, the flow line generation unit 126 may set the position of the first object as the predicted position in the second time to the fixed position of the object in the second time by modifying the position of the first object based on the candidate position selected by the selection unit 125. .. FIG. 13 is a diagram showing an example in which a fixed position corresponding to the second time is specified. In FIG. 13, the fixed position PFA at the second time t is specified based on the relationship between the first object position PPA and the candidate position P2A, and based on the relationship between the first object position PPB and the candidate position P2B, It can be confirmed that the fixed position PFB at the second time t is specified. By doing so, the flow line generation device 1 can accurately specify the position of the object at the second time t.

[Process flow in the flow line generator 1]
Subsequently, the flow of processing in the flow line generation device 1 will be described. FIG. 14 is a flowchart showing a processing flow in the flow line generation device 1 according to the present embodiment.

First, the obstacle area specifying unit 121 identifies an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area (S1).
Subsequently, the object position specifying unit 122 specifies the positions of one or more objects moving in the predetermined area before the first time (S2). Specifically, the object position specifying unit 122 specifies one or more flow line information including the position information up to the first time or the position information up to the time before the first time, and the latest time in the flow line information. By specifying the position information associated with, the position of one or more objects before the first time is specified.

Subsequently, the detection position specifying unit 123 identifies a candidate position that is a candidate for one or more detection positions of the object at the second time, which is a time after the first time (S3).
Subsequently, the movement obstruction degree specifying unit 124 specifies the movement obstruction degree for each of the one or more candidate positions of one or more first object positions (S4).

Subsequently, the selection unit 125 specifies the degree of similarity of one or more objects with the objects corresponding to each of the one or more candidate positions (S5).
Subsequently, the selection unit 125 identifies the movement cost corresponding to each of the one or more candidate positions of one or more objects based on the movement obstruction degree specified in S4 and the similarity specified in S5. (S6).

Subsequently, the selection unit 125 corresponds to one or more objects and one or more candidate positions based on the movement cost corresponding to each of the one or more candidate positions of one or more objects specified in S6. Identify the relationship (S7).

Subsequently, the flow line generation unit 126 sets the second time and the position information indicated by the candidate position specified for the object indicated by the flow line information to each of the one or more flow line information specified in S2. The flow line information is updated by associating and storing it (S8).

[Effects in this embodiment]
As described above, the flow line generation device 1 according to the present embodiment specifies the position of the obstacle area corresponding to the predetermined area in the predetermined area, and specifies the position of the object moving in the predetermined area before the first time. At the same time, one or more candidate positions of the object at the second time, which is a time after the first time, are specified. In the flow line generator 1, the object is one or more candidate positions based on the positional relationship between the specified position of the object before the first time, the specified one or more candidate positions, and the specified position of the obstacle area. Specify the degree of obstruction when moving to each of the above. Then, the flow line generation device 1 specifies the movement cost of the object from the position before the first time to each of the one or more candidate positions based on the degree of movement obstruction, and selects the candidate position where the movement cost is relatively small. , Select the position of the object at the second time. By doing so, the flow line generation device 1 can efficiently specify the correspondence between the objects.

Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various changes or improvements can be made to the above embodiments.

Further, in particular, the specific embodiment of the distribution / integration of the apparatus is not limited to those shown above, and all or a part thereof may be arbitrary according to various additions or the like, or according to the functional load. It can be functionally or physically distributed / integrated in units.

1 ... Flow line generator, 11 ... Storage unit, 12 ... Control unit, 121 ... Obstacle area identification unit, 122 ... Object position identification unit, 123 ... Detection position identification unit, 124 ... Movement obstruction degree specifying part, 125 ... Selection part, 126 ... Flow line generation part

Claims (21)

An obstacle area corresponding to the shape of an obstacle existing in a predetermined area, an obstacle area specifying portion for specifying a position of the obstacle area in the predetermined area, and an obstacle area specifying portion.
An object position specifying unit that specifies the positions of a plurality of objects moving in the predetermined area before the first time, and
A detection position specifying unit that identifies a plurality of candidate positions that are candidates for a plurality of detection positions corresponding to each of the plurality of objects at the second time, which is a time after the first time.
For each of the positions of the plurality of objects before the first time specified, based on the positional relationship between the position of the object, the specified plurality of candidate positions, and the position of the specified obstacle area. The movement when each of the plurality of objects moves to each of the plurality of candidate positions by specifying the degree of movement obstruction indicating the difficulty when the object moves to each of the plurality of candidate positions. The movement obstruction degree specification part that specifies the obstruction degree, and the movement obstruction degree identification part,
The element is a cost value indicating the movement cost based on the movement obstruction degree when the object position specifying unit moves from the position of each of the plurality of objects before the first time to each of the plurality of candidate positions. By generating a cost matrix to be generated and applying the Hungarian method to the generated cost matrix, the cost when the object is moved from the position of each of the plurality of objects before the first time to one of the plurality of candidate positions. A plurality of combinations of the object and the candidate position having a relatively small sum of values are specified so that the candidate positions do not overlap, and the candidate positions corresponding to each of the plurality of objects indicated by the specified combination are specified. , A selection unit selected at the position of each of the plurality of objects at the second time,
A position selection device comprising. The object position specifying unit predicts the position of the first object, which is the position of the object at the second time, based on the position of the object before the first time.
The movement obstruction degree specifying unit is based on the positional relationship between the first object position predicted for the object, the plurality of candidate positions specified for the object, and the position of the specified obstacle region. To identify the movement obstruction degree corresponding to the object,
The position selection device according to claim 1. The object position specifying unit specifies the position of the object at the first time as the first object position.
The movement obstruction degree specifying unit attaches to the object based on the positional relationship between the position of the first object specified for the object, the specified plurality of candidate positions, and the position of the specified obstacle area. Identify the corresponding movement obstruction,
The position selection device according to claim 1. The movement obstruction degree specifying unit specifies a line segment connecting the first object position corresponding to the object and each of the specified plurality of candidate positions, and the line segment and the specified obstacle area. Based on the positional relationship, the movement obstruction degree corresponding to each of the specified plurality of candidate positions is specified for the object.
The position selection device according to claim 2 or 3. The movement obstruction degree specifying unit identifies two intersections of the line segment and the specified obstacle area for each of the specified plurality of line segments, and the specified two intersections and the obstacle area. To identify the movement obstruction degree corresponding to the object, based on the relationship of
The position selection device according to claim 4. The movement obstruction degree specifying unit specifies the movement obstruction degree based on the length of a relatively short contour line among the two contour lines of the obstacle region divided by the two specified intersections.
The position selection device according to claim 5. The movement obstruction degree specifying unit can be taken from among the three intersections when there are three or more intersections between the line segment and the specified obstacle region for each of the specified plurality of the line segments. Identify two intersections with relatively long distances,
The position selection device according to claim 5 or 6. The movement obstruction degree specifying unit is based on the ratio of the length of the contour line of the obstacle region divided by the two specified intersections to the length of the line segment connecting the two specified intersections. To identify,
The position selection device according to any one of claims 5 to 7. The movement obstruction degree specifying unit specifies the angle formed by each of the two specified intersections and the two line segments connecting the specified obstacle region reference point, and determines the movement obstruction degree based on the angle. Identify,
The position selection device according to any one of claims 5 to 8. The movement obstruction degree specifying unit identifies two sides in which any of the two specified intersections exists among the sides constituting the specified obstacle area, and two side groups existing between the two sides. Of these, the degree of movement obstruction is specified based on the number of sides included in the side group having a relatively small number of sides.
The position selection device according to any one of claims 5 to 9. When there are a plurality of obstacle regions having intersections with the specified line segments, the movement obstruction degree specifying unit approximates the plurality of obstacle regions to one obstacle region.
The position selection device according to any one of claims 5 to 10. When the first object position exists in the obstacle area, the object position specifying unit corrects the first object position to a position outside the obstacle area.
The position selection device according to any one of claims 2 to 11. When the shape of the obstacle is non-convex, the obstacle region specifying portion approximates the obstacle region corresponding to the obstacle to the convex obstacle region.
The position selection device according to any one of claims 1 to 12. When a plurality of the specified obstacle areas are specified and the specified plurality of obstacle areas are adjacent to each other or overlapped with each other, the obstacle area specifying unit can use the plurality of obstacle areas as the plurality of obstacle areas. Approximate to the sum region of
The position selection device according to any one of claims 1 to 13. When the specified candidate position exists in the obstacle area, the detection position specifying unit corrects the candidate position to a position outside the obstacle area.
The position selection device according to any one of claims 1 to 14. The selection unit specifies the degree of similarity in a predetermined correspondence between the specified object corresponding to the position before the first time and the object corresponding to each of the specified plurality of candidate positions, and the degree of similarity and the object. Based on the identified movement obstruction degree, the movement cost from the position before the first time to each of the plurality of candidate positions is specified.
The position selection device according to any one of claims 1 to 15. The selection unit controls so that the candidate position is not selected when the relatively small movement cost exceeds the first threshold value.
The position selection device according to any one of claims 1 to 16. The selection unit sets the movement cost to a value exceeding the first threshold value for the candidate position whose specified movement interference degree exceeds the second threshold value.
The position selection device according to claim 17. A flow line generation unit that generates a movement flow line of the object by associating the position of the object at the first time with the candidate position of the object selected by the selection unit is further provided.
The position selection device according to any one of claims 1 to 18. Computer runs,
A step of specifying an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area.
A step of identifying the positions of a plurality of objects moving in the predetermined area before the first time, and
A step of specifying a plurality of candidate positions that are candidates for a plurality of detection positions corresponding to each of the plurality of objects at the second time, which is a time after the first time,
For each of the positions of the plurality of objects before the first time specified, based on the positional relationship between the position of the object, the specified plurality of candidate positions, and the position of the specified obstacle area. The movement when each of the plurality of objects moves to each of the plurality of candidate positions by specifying the degree of movement obstruction indicating the difficulty when the object moves to each of the plurality of candidate positions. Steps to identify the degree of interference and
From the positions of the plurality of objects before the first time specified, a cost matrix having a cost value indicating the movement cost based on the movement obstruction degree when moving to each of the plurality of candidate positions is generated as an element. Steps to do and
By applying the Hungarian method to the generated cost matrix, the sum of the cost values when moving from the position of each of the plurality of objects before the first time to any of the plurality of candidate positions is relative. A plurality of small combinations of the object and the candidate position are specified so that the candidate positions do not overlap, and the candidate positions corresponding to each of the plurality of objects indicated by the specified combination are set at the second time. Steps to select at each position of the multiple objects,
Position selection method. Computer,
An obstacle area specifying portion that specifies an obstacle area corresponding to the shape of an obstacle existing in a predetermined area and a position of the obstacle area in the predetermined area.
An object position specifying unit that specifies the positions of a plurality of objects moving in the predetermined area before the first time.
A detection position specifying unit that identifies a plurality of candidate positions that are candidates for a plurality of detection positions corresponding to each of the plurality of objects at the second time, which is a time after the first time.
For each of the positions of the plurality of objects before the first time specified, based on the positional relationship between the position of the object, the specified plurality of candidate positions, and the position of the specified obstacle area. The movement when each of the plurality of objects moves to each of the plurality of candidate positions by specifying the degree of movement obstruction indicating the difficulty when the object moves to each of the plurality of candidate positions. Movement obstruction degree specification part that specifies the obstruction degree, and
The element is a cost value indicating the movement cost based on the movement obstruction degree when the object position specifying unit moves from the position of each of the plurality of objects before the first time to each of the plurality of candidate positions. By generating a cost matrix to be generated and applying the Hungarian method to the generated cost matrix, the cost when the object is moved from the position of each of the plurality of objects before the first time to one of the plurality of candidate positions. A plurality of combinations of the object and the candidate position having a relatively small sum of values are specified so that the candidate positions do not overlap, and the candidate positions corresponding to each of the plurality of objects indicated by the specified combination are specified. , A selection unit selected at the position of each of the plurality of objects at the second time,
A position selection program that functions as.

Images