JP6937266B2 - Area selection device, area selection method, and area selection program - Google Patents

Description

The present invention relates to an area selection device for matching rectangular areas corresponding to objects, an area selection method, and an area selection program.

Conventionally, an image captured by an imaging device installed in a predetermined area is analyzed to identify an object area, and a storage unit that associates the object area with a position in the predetermined area of the object is referred to and specified. An image analysis device that identifies a position of an object in a predetermined area is disclosed (see, for example, Patent Document 1).

JP-A-2017-117244

By the way, when matching the area of the specified object with the object stored in the storage unit, the overlap rate (IOU: Intersection over Union) between the area of the specified object and the object stored in the storage unit is set. Matching is done based on. However, in the case of matching based on the duplication rate, it is necessary to calculate the duplication rate between the area of the specified object and all the objects stored in the storage unit, and there is a problem that matching takes time.

On the other hand, in Patent Document 1, the center coordinates and area of the indexed object and the position of the object in a predetermined area are stored in association with each other, and the center coordinates and area of the specified object and the stored object are stored. It is disclosed that matching with the center coordinates and area of the object is performed at high speed.

However, when indexing is performed in a three-dimensional space in which the center coordinates and the area are defined as one vector, there is a problem that the accuracy is inferior to that in the case of matching using the overlap rate.

Therefore, the present invention has been made in view of these points, and an area selection device, an area selection method, and an area selection program capable of realizing matching of objects corresponding to a specified object at high speed and with high accuracy are provided. The purpose is to provide.

The area selection device according to the first aspect of the present invention identifies an object included in the image captured by the image pickup device that images a predetermined area, and determines the position of the object corresponding to the local position which is the position in the image. It is included in the related information in which the rectangular area specifying portion that specifies the position and size of the rectangular area corresponding to the object and the rectangular area information indicating the position and size of the rectangular area are associated with each of the plurality of local positions. A first selection unit that selects one or more local positions that are relatively close to the position of the object specified by the rectangular area identification unit from the plurality of local positions, and the rectangular area specified by the rectangular area identification unit. The one or more said, based on the position and size and the position and size of the rectangular area indicated by the one or more rectangular area information associated with each of the one or more local positions selected by the first selection. A second selection unit for selecting one rectangular area information from the rectangular area information is provided.

In the related information, each of the plurality of local positions is further associated with a global position which is a position in the predetermined area of the object corresponding to the rectangular area, and the area selection device is the first in the related information. (2) A position specifying unit that specifies the global position associated with the one rectangular area information selected by the selection unit as the global position of the object may be further provided.

The second selection unit is a rectangular area indicated by the rectangular area specified by the rectangular area identification unit and one or more rectangular area information associated with each of the one or more local positions selected by the first selection unit. The overlap rate with each of the above may be calculated, and one of the rectangular area information corresponding to the rectangular area having a relatively high overlap rate may be selected.

The second selection unit may control not to select one of the rectangular area information when the highest overlap rate among the calculated overlap rates is less than a predetermined threshold value.

The second selection unit does not select one of the rectangular area information when the local position selected by the first selection unit and the position of the object specified by the rectangular area identification unit are equal to or longer than a predetermined distance. You may control it.

The first selection unit may control not to select the local position when the related information does not include a local position within a predetermined distance from the position of the object specified by the rectangular area identification unit. ..

The related information is provided corresponding to each of the plurality of shapes that the object can take, and the first selection unit refers to the plurality of related information corresponding to each of the plurality of shapes, and the rectangle. A plurality of local positions corresponding to the positions of the objects specified by the area specifying unit may be selected.

The first selection unit refers to the plurality of related information and selects a plurality of the local positions corresponding to each of the plurality of shapes, and the second selection unit has the first selection unit having a plurality of shapes. Based on the position and size of the rectangular area indicated by the rectangular area information associated with the plurality of local positions selected for each, and the position and size of the rectangular area specified by the rectangular area specifying portion. One rectangular area information may be selected from the plurality of rectangular area information selected for each of the plurality of shapes.
In the relevant information, the plurality of local locations may be indexed.

The area selection method according to the second aspect of the present invention identifies an object included in an image captured by an imaging device that captures a predetermined area, which is executed by a computer, and corresponds to a local position which is a position in the image. The step of specifying the position of the object and the position and size of the rectangular area corresponding to the object, and the related information in which the rectangular area information indicating the position and size of the rectangular area is associated with each of the plurality of local positions. A step of selecting one or more local positions relatively close to the specified object position from the plurality of included local positions, the position and size of the specified rectangular area, and one or more selected local positions. A step of selecting one rectangular area information from the one or more rectangular area information based on the position and size of the rectangular area indicated by the one or more rectangular area information associated with each of the local positions. Be prepared.

The area selection program according to the third aspect of the present invention identifies an object included in an image captured by an imaging device that images a predetermined area, and the object corresponds to a local position that is a position in the image. Related information in which the rectangular area specifying part that specifies the position and the position and size of the rectangular area corresponding to the object and the rectangular area information indicating the position and size of the rectangular area are associated with each of the plurality of local positions. A first selection unit that selects one or more local positions that are relatively close to the position of the object specified by the rectangular area identification unit from the plurality of included local positions, and the rectangular area identification unit that identifies the first selection unit. Based on the position and size of the rectangular area and the position and size of the rectangular area indicated by the one or more rectangular area information associated with each of the one or more local positions selected by the first selection. It functions as a second selection unit for selecting one rectangular area information from the above rectangular area information.

According to the present invention, there is an effect that matching of an object corresponding to a specified object can be realized at high speed and with high accuracy.

It is a figure which shows the outline of the area selection apparatus which concerns on 1st Embodiment. It is a figure which shows the structure of the area selection apparatus which concerns on 1st Embodiment. It is a figure which shows an example of the object projection information which concerns on 1st Embodiment. It is a figure explaining the example of selecting a plurality of local positions. It is a flowchart which shows the flow of processing in the area selection apparatus which concerns on 1st Embodiment. It is a figure which shows the structure of the area selection apparatus which concerns on 2nd Embodiment. It is a figure which shows an example of the object projection information which concerns on 2nd Embodiment.

<First Embodiment>
[Overview of area selection device 2]
FIG. 1 is a diagram showing an outline of the area selection device 2 according to the first embodiment. The area selection device 2 is a device that identifies a rectangular area corresponding to an object included in an image captured by the image pickup device 1. The object is, for example, a person such as a clerk who is a target for tracking a movement trajectory in a predetermined area such as a store.

The image pickup device 1 is an image pickup device that images a predetermined area. Here, the predetermined area is, for example, the floor of a store. In the predetermined area, a person such as a clerk exists as an object for tracking the trajectory of movement. The image pickup device 1 captures an omnidirectional image of a predetermined area. In the present embodiment, the image pickup device 1 captures an omnidirectional image, but the present invention is not limited to this, and an image other than the omnidirectional image may be captured.

The area selection device 2 acquires an omnidirectional image captured by the image pickup device 1 ((1) in FIG. 1). The area selection device 2 analyzes the acquired omnidirectional image to specify a rectangular area indicating an object included in the omnidirectional image and to specify the position of the object on the omnidirectional image ((2) in FIG. 1). )). At this time, the area selection device 2 can use a machine learner or a deep learner that has been trained for object detection.

The area selection device 2 includes rectangular area information indicating a rectangular area corresponding to the object included in the omnidirectional image captured by the image pickup device 1 when the object is arranged at the global position which is the position of the predetermined area, and the said area selection device 2. Object projection information is stored as related information associated with the local position, which is the position on the omnidirectional image of the object. The area selection device 2 refers to the object projection information and selects one or more local positions close to the position of the specified object ((3) in FIG. 1).

The area selection device 2 calculates the overlap rate between the rectangular area of the specified object and the rectangular area associated with each of the selected one or more local positions ((4) in FIG. 1), and the overlap rate is determined. The rectangular area information indicating the highest rectangular area is specified ((5) in FIG. 1).

By doing so, the area selection device 2 does not calculate the overlap rate between the rectangular area indicating the object specified from the omnidirectional image and all the rectangular areas stored in the object projection information. Since the rectangular area information is specified, it is possible to match the rectangular area indicating the object with high speed and high accuracy as compared with the case of calculating the overlap rate with all the rectangular areas.
Subsequently, the configuration of the area selection device 2 will be described.

[Configuration of area selection device 2]
FIG. 2 is a diagram showing a configuration of the area selection device 2 according to the first embodiment. As shown in FIG. 2, the area selection device 2 includes a storage unit 21 and a control unit 22.

The storage unit 21 is a storage medium including a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The storage unit 21 stores a program executed by the control unit 22. For example, the storage unit 21 stores an area selection program that causes the area selection device 2 to function as an image acquisition unit 221, a rectangular area identification unit 222, a first selection unit 223, and a second selection unit 224.

Further, it is assumed that the storage unit 21 stores in advance object projection information in which rectangular area information indicating the position and size of the rectangular area is associated with each of the plurality of local positions.

FIG. 3 is a diagram showing an example of object projection information according to the first embodiment. As shown in FIG. 3, the object projection information is provided corresponding to each of a plurality of shapes that the object can take. Since the object is a person, the shape of the object is, for example, the shape of a person when standing upright, the shape of a person when sitting on a chair, the shape of a person indicating an adult, and the shape of a person indicating a child. Is. The object projection information may be provided separately for each of the plurality of shapes, or may be provided only once for each of the plurality of shapes.

Further, in the object projection information, a plurality of local positions are indexed by, for example, a kd tree (k-dimensional tree). Since the local position is indexed, the search for the local position can be realized at high speed.

The control unit 22 is, for example, a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). The control unit 22 functions as an image acquisition unit 221, a rectangular area identification unit 222, a first selection unit 223, and a second selection unit 224 by executing an area selection program stored in the storage unit 21.

The image acquisition unit 221 acquires an omnidirectional image captured by the image pickup device 1. The omnidirectional image contains information indicating the imaging time.

The rectangular area specifying unit 222 identifies the position and size of the rectangular area corresponding to the object in the omnidirectional image by analyzing the omnidirectional image captured by the imaging device 1 that images the predetermined area. For example, the rectangular area specifying unit 222 specifies two-dimensional coordinates (x, y) indicating one vertex of the rectangular area as the position of the rectangular area in the omnidirectional image, and determines the vertex as the size of the rectangular area. By specifying the distances w and h between the two vertices adjacent to each other, the rectangular area information (x, y, w, h) indicating the position and size of the rectangular area is specified. Here, it is assumed that the inclination θ of the rectangular region with respect to the predetermined axis on the omnidirectional image is 0. A rectangular area with an inclination θ of 0 is called an upright rectangular area. The rectangular area specifying unit 222 may specify the rectangular area of the object by using a machine learning device or a deep learning device. In general, when a rectangular area of an object is specified using a machine learner or a deep learner, it is often specified as an upright rectangular area.

Further, the rectangular area specifying unit 222 specifies the position of the object in the omnidirectional image based on the position and size of the rectangular area corresponding to the specified object. For example, the rectangular area specifying unit 222 specifies the center point of the rectangular area as the position of the object in the omnidirectional image. The position of the object specified here does not have to be the same as the two-dimensional coordinates (x, y) included in the rectangular area information.

The first selection unit 223 selects one or more local positions relatively close to the position of the object specified by the rectangular area identification unit 222 from a plurality of local positions included in the object projection information stored in the storage unit 21. do.

Specifically, the first selection unit 223 refers to a plurality of object projection information corresponding to each of the plurality of shapes that the object can take, and the object specified by the rectangular area specifying unit 222 with respect to the plurality of object projection information. As a local position relatively close to the position of, one or more local positions within a predetermined distance from the position of the object are selected.

FIG. 4 is a diagram illustrating an example of selecting a plurality of local positions. The rectangular area R shown in FIG. 4 is a rectangular area corresponding to the object specified by the rectangular area specifying unit 222 from the omnidirectional image OP, and the position P is the position of the object specified by the rectangular area specifying unit 222.

The first selection unit 223 refers to the object projection information corresponding to the upright state, and selects one or more local positions as the local positions corresponding to the range AP within a predetermined distance from the position P. Further, the first selection unit 223 refers to the object projection information corresponding to the seated state, and selects one or more local positions as the local positions corresponding to the range AP within a predetermined distance from the position P. Since the plurality of local positions included in the object projection information are indexed, the first selection unit 223 can select one or more local positions at high speed.

Note that the first selection unit 223 does not select a local position from the object projection information when the object projection information does not include a local position within a predetermined distance from the position of the object specified by the rectangular area identification unit 222. May be good. By doing so, the area selection device 2 can suppress over-detection in which the rectangular area information is selected even though the rectangular area information corresponding to the object does not exist.

The second selection unit 224 contains information on one or more rectangular areas associated with the position and size of the rectangular area specified by the rectangular area identification unit 222 and one or more local positions selected by the first selection unit 223. One rectangular area information is selected from the one or more rectangular area information based on the position and size of the indicated rectangular area.

Specifically, the second selection unit 224 is one or more associated with the rectangular area indicated by the rectangular area information specified by the rectangular area identification unit 222 and one or more local positions selected by the first selection unit 223. The overlap rate with each of the rectangular areas indicated by the rectangular area information is calculated. Then, the second selection unit 224 selects one rectangular area information corresponding to the rectangular area having a relatively high overlap rate.

More specifically, the second selection unit 224 refers to the position and size of the rectangular area indicated by the rectangular area information associated with the plurality of local positions selected by the first selection unit 223 for each of the plurality of shapes. , Select one rectangular area information from the rectangular area information selected for each of a plurality of shapes based on the overlap rate calculated based on the position and size of the rectangular area specified by the rectangular area specifying unit 222. .. For example, the second selection unit 224 selects one rectangular area information having the highest duplication rate among the rectangular area information selected for each of the plurality of shapes.

When the first selection unit 223 selects a local position relatively close to the object position from all the rectangular area information, only the local position corresponding to the specific shape is selected and corresponds to the actual object shape. You may not be able to select the local location to do. On the other hand, in the area selection device 2, the first selection unit 223 selects one rectangular area information from the rectangular area information associated with the plurality of local positions selected for each of the plurality of shapes. The rectangular area information corresponding to the shape of the actual object can be selected with high accuracy.

The second selection unit 224 may control not to select one rectangular area information when the highest overlap rate among the calculated plurality of overlap rates is less than a predetermined threshold value. Further, the second selection unit 224 does not select one rectangular area information when the local position selected by the first selection unit 223 and the position of the object specified by the rectangular area identification unit 222 are equal to or longer than a predetermined distance. You may. By doing so, the region selection device 2 can suppress over-detection.

[Process flow in the area selection device 2]
Subsequently, the processing flow in the area selection device 2 will be described. FIG. 5 is a flowchart showing a processing flow in the area selection device 2 according to the first embodiment.

First, the image acquisition unit 221 acquires an omnidirectional image from the image pickup device 1 (S11).
Subsequently, the rectangular area specifying unit 222 performs image analysis of the acquired omnidirectional image, and specifies the rectangular area of the object included in the omnidirectional image and the position of the object (S12).

Subsequently, the first selection unit 223 starts from a plurality of local positions included in a plurality of object projection information corresponding to each of the plurality of shapes that the object stored in the storage unit 21 can take, and from the position of the specified object. Select one or more local positions that are relatively close (S13).

Subsequently, the second selection unit 224 includes a rectangular area of the specified object and a rectangular area indicated by the rectangular area information associated with each of the one or more local positions corresponding to each of the plurality of selected shapes. The duplication rate of is calculated (S14).
Subsequently, the second selection unit 224 selects the rectangular area information indicating the rectangular area having the highest calculated duplication rate (S15).

[Effect in the first embodiment]
As described above, the area selection device 2 according to the first embodiment identifies an object included in the image captured by the image pickup device 1 that images a predetermined area, and the position of the object corresponding to the local position and the said object. Identify the position and size of the rectangular area corresponding to the object. The area selection device 2 selects one or more local positions relatively close to the position of the specified object from the object projection information, and in the position and size of the rectangular area corresponding to the specified object and the object projection information. Select one rectangular area information from the one or more rectangular area information based on the position and size of the rectangular area indicated by the one or more rectangular area information associated with each of the selected one or more local positions. .. By doing so, the area selection device 2 can realize matching between the object specified from the omnidirectional image and the object indicated by the rectangular area information on the object projection information at high speed and with high accuracy.

<Second Embodiment>
[Select rectangular area information corresponding to any of the plurality of image pickup devices 1]
Subsequently, the second embodiment will be described. In the first embodiment, the object projection information corresponding to the image pickup device 1 is referred to, and one rectangular area information corresponding to the object included in the omnidirectional image captured by the image pickup device 1 is selected. However, depending on the size of the predetermined area, one imaging device 1 may not be able to image the entire predetermined area, and a plurality of imaging devices 1 may image the predetermined area. Therefore, the area selection device 2 according to the second embodiment selects one rectangular area information from the omnidirectional images captured by the plurality of image pickup devices 1.

FIG. 6 is a diagram showing a configuration of the area selection device 2 according to the second embodiment. As shown in FIG. 6, the control unit 22 of the area selection device 2 further includes a position specifying unit 225.
FIG. 7 is a diagram showing an example of object projection information according to the second embodiment. As shown in FIG. 7, the object projection information according to the second embodiment stores the device ID that identifies each of the plurality of imaging devices 1, the local position, the rectangular area information, and the global position in association with each other.

In the second embodiment, the first selection unit 223 refers to a plurality of object projection information corresponding to each of the plurality of shapes that the object can take, and is a source for specifying the position of the object with respect to the plurality of object projection information. A plurality of local positions corresponding to the position of the object are selected from the plurality of local positions associated with the device ID of the imaging device that has captured the omnidirectional image.

For example, the first selection unit 223 has a plurality of locals associated with a device ID indicating the first image pickup device 1 for the position of the specified object with respect to the omnidirectional image acquired from the first image pickup device 1. From the positions, select a plurality of local positions corresponding to the positions of the object. Similarly, the first selection unit 223 is associated with a plurality of device IDs indicating the second image pickup device 1 for the position of the specified object with respect to the omnidirectional image acquired from the second image pickup device 1. From the local positions, select a plurality of local positions corresponding to the position of the object.

Since the first image pickup device 1 and the second image pickup device 1 simultaneously image a predetermined area, if the image pickup ranges of the first image pickup device 1 and the second image pickup device 1 overlap, they respectively. The object may be reflected in the omnidirectional image captured by the image pickup device 1 of the above. In this case, the first selection unit 223 selects a plurality of local positions associated with the device ID indicating the first image pickup device 1, and a plurality of positions associated with the device ID indicating the second image pickup device 1. Select a local location for.

The second selection unit 224 contains information on one or more rectangular areas associated with the position and size of the rectangular area specified by the rectangular area identification unit 222 and one or more local positions selected by the first selection unit 223. One rectangular area information is selected from the one or more rectangular area information based on the position and size of the indicated rectangular area. The second selection unit 224 selects a plurality of local positions associated with the device ID indicating the first image pickup device 1 with respect to one time, and the second selection unit 224 selects a plurality of local positions associated with the device ID, and the second image pickup device 1 When a plurality of local positions associated with the device ID indicating the device ID are selected, one rectangular area information may be selected for each device ID. In this case, the second selection unit 224 selects two rectangular area information.

The position specifying unit 225 is an object in which the rectangular area specifying unit 222 specifies the global position associated with the one rectangular area information selected by the second selection unit 224 in the object projection information stored in the storage unit 21. Identify as the global location of. Further, in the position specifying unit 225, when the second selection unit 224 selects two rectangular area information, the rectangular area specifying unit 222 specifies, for example, the global position associated with the rectangular area information having the higher duplication rate. Specify as the global position of the created object. When the second selection unit 224 selects two rectangular area information, the position specifying unit 225 sets the global position associated with each rectangular area information as the global position of the object specified by the rectangular area specifying unit 222. May be specified as.

The position specifying unit 225 associates the information indicating the imaging time included in the omnidirectional image including the specified object with the position information indicating the specified global position, and stores the object position information as the storage unit 21. To memorize.

[Effect in the second embodiment]
As described above, the area selection device 2 according to the second embodiment specifies the global position associated with the specified rectangular area information to the global position of the specified object. By doing so, the area selection device 2 can specify the global position corresponding to the object included in at least one of the omnidirectional images captured by the plurality of image pickup devices 1 at high speed and with high accuracy.

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 arbitrarily added according to various additions or functional loads. It can be functionally or physically distributed / integrated in units.

1 ... Imaging device, 2 ... Area selection device, 21 ... Storage unit, 22 ... Control unit, 221 ... Image acquisition unit, 222 ... Rectangular area identification unit, 223 ... 1st selection unit, 224 ... 2nd selection unit, 225 ... Position identification unit

Claims (11)

The object included in the image captured by the imaging device that captures the predetermined area is specified, and the position of the object corresponding to the local position, which is the position in the image, and the position and size of the rectangular area corresponding to the object are specified. Rectangle area identification part and
Relative to the position of the object specified by the rectangular area specifying portion from the plurality of local positions included in the related information in which the rectangular area information indicating the position and size of the rectangular area is associated with each of the plurality of local positions. A first selection that selects one or more local locations that are close to each other,
The position and size of the rectangular area specified by the rectangular area specifying unit and the position of the rectangular area indicated by one or more rectangular area information associated with each of the one or more local positions selected by the first selection unit. A second selection unit that selects one rectangular area information from the one or more rectangular area information based on the size and the size.
Area selection device comprising. In the related information, each of the plurality of local positions is further associated with a global position, which is a position in the predetermined area of the object corresponding to the rectangular area.
A position specifying unit that specifies the global position associated with the one rectangular area information selected by the second selection unit as the global position of the object in the related information is further provided.
The area selection device according to claim 1. The second selection unit is a rectangular area indicated by the rectangular area specified by the rectangular area identification unit and one or more rectangular area information associated with each of the one or more local positions selected by the first selection unit. The overlap rate with each of the above is calculated, and one of the rectangular area information corresponding to the rectangular area having a relatively high overlap rate is selected.
The area selection device according to claim 1. The second selection unit controls not to select one of the rectangular area information when the highest overlap rate among the calculated overlap rates is less than a predetermined threshold value.
The area selection device according to claim 3. The second selection unit does not select one of the rectangular area information when the local position selected by the first selection unit and the position of the object specified by the rectangular area identification unit are equal to or longer than a predetermined distance. Control,
The area selection device according to any one of claims 1 to 4. The first selection unit controls not to select the local position when the related information does not include a local position within a predetermined distance from the position of the object specified by the rectangular area specifying unit.
The area selection device according to any one of claims 1 to 5. The related information is provided corresponding to each of a plurality of shapes that the object can take.
The first selection unit refers to a plurality of the related information corresponding to each of the plurality of shapes, and selects a plurality of local positions corresponding to the positions of the objects specified by the rectangular area identification unit.
The area selection device according to any one of claims 1 to 6. The first selection unit refers to a plurality of the related information, selects a plurality of the local positions corresponding to each of the plurality of shapes, and selects a plurality of the local positions.
The second selection unit includes the position and size of the rectangular area indicated by the rectangular area information associated with the plurality of local positions selected by the first selection unit for each of the plurality of shapes, and the rectangular area. One rectangular area information is selected from a plurality of the rectangular area information selected for each of the plurality of shapes based on the position and size of the rectangular area specified by the specific portion.
The area selection device according to claim 7. In the relevant information, the plurality of local locations are indexed.
The area selection device according to any one of claims 1 to 8. Computer runs,
The object included in the image captured by the imaging device that captures the predetermined area is specified, and the position of the object corresponding to the local position, which is the position in the image, and the position and size of the rectangular area corresponding to the object are specified. Steps to do and
One or more that are relatively close to the position of the object specified from the plurality of local positions included in the related information in which the rectangular area information indicating the position and size of the rectangular area is associated with each of the plurality of local positions. Steps to select a local location for
One or more based on the identified position and size of the rectangular area and the position and size of the rectangular area indicated by one or more of the rectangular area information associated with each of the selected one or more local positions. And the step of selecting one rectangular area information from the rectangular area information of
Area selection method including. Computer,
The object included in the image captured by the imaging device that captures the predetermined area is specified, and the position of the object corresponding to the local position, which is the position in the image, and the position and size of the rectangular area corresponding to the object are specified. Rectangular area identification part,
Relative to the position of the object specified by the rectangular area specifying portion from the plurality of local positions included in the related information in which the rectangular area information indicating the position and size of the rectangular area is associated with each of the plurality of local positions. A first selection that selects one or more local locations that are close to each other, and
The position and size of the rectangular area specified by the rectangular area specifying unit and the position of the rectangular area indicated by one or more rectangular area information associated with each of the one or more local positions selected by the first selection unit. A second selection unit that selects one rectangular area information from the one or more rectangular area information based on the size and the size.
Area selection program to function as.

Images