JP7185740B1 - Area identification device, area identification method, and area identification program - Google Patents

Abstract

A cross-sectional area in image data can be specified without human intervention. An object detection unit (23) detects an object of each of two or more types as a target type from image data obtained by a photographing device. The connotative region identifying unit 241 identifies the connotative region containing the position of the object for the type of the detected target. The overlapping region identification unit 242 identifies an overlapping region in which the connotative region identified for one type and the connotative region identified for the other type among the two or more types overlap, and an area in which an object of a certain type moves. is identified as a traversing area traversed by an object of the type [Selection drawing] Fig. 1

Description

The present disclosure relates to a technique for identifying a crossing area in which another object traverses a moving area of a certain type of object.

When a person or vehicle is traveling, there is a no-entry area that is temporarily prohibited from entering due to the movement of other objects. Although there is a gate or the like to prevent entry into this prohibited area, there are cases where people or vehicles enter and cause accidents. In addition, there are cases in which people cross the pedestrian crossing even though the traffic light is red, leading to an accident. In order to prevent accidents, crossing areas such as no-entry areas and pedestrian crossings are monitored using video data obtained by surveillance cameras.

When monitoring using video data, setting a monitoring target area and limiting detection to the set area is effective in terms of suppressing false detections and improving detection performance. .
The positions and angles installed by the surveillance cameras are different, and the size and shape of the areas to be monitored are different. Therefore, setting of the monitoring target area is generally performed manually. Therefore, when the number of monitoring cameras increases, setting the area to be monitored becomes a heavy burden. In addition, if the monitoring camera moves due to the influence of the wind or the like and the imaging range changes, it is necessary to reset the area to be monitored, which causes a burden.

Japanese Patent Laid-Open No. 2002-200002 describes determining whether or not a traffic violation has occurred by determining whether or not a vehicle block is in a restricted area in a captured image.

JP 2007-316856 A

Japanese Patent Application Laid-Open No. 2002-300000 also does not describe a method for setting the restricted area, and it is necessary to set it manually. Therefore, even when determining whether or not there is a traffic violation using the technique described in Patent Document 1, a burden is placed on setting the restricted area.
An object of the present disclosure is to enable identification of a crossing region in image data without manual intervention.

The area identification device according to the present disclosure is
an object detection unit that detects an object of the target type from image data obtained by a photographing device, with each of two or more types as a target type;
an enclosing area specifying unit that specifies an enclosing area that encloses the position of the object for the type of the target detected by the object detecting unit;
A region where an object of a certain type moves in a superimposed region where the connotative region identified for one type among the two or more types by the connotative region identifying unit overlaps the connotative region identified for another type. as a traversing region traversed by another type of object.

The encapsulating area specifying unit specifies the enclosing area that includes the positions remaining after outliers are removed from the positions of the object for the type of the object detected by the object detecting unit.

The object detection unit tracks the detected object,
The enclosing area identifying unit identifies the enclosing area that includes the remaining positions after removing the positions of the object that could not be tracked.

The area identification device further
A target detection unit that detects a detection target from the crossing area specified by the superimposed area specifying unit in the image data.

After detecting the object, the object detection unit plots a point at a position set by the user for each type of object,
The encapsulation area specifying unit specifies the encapsulation area based on the plotted points.

The object detection unit also detects an object other than the object of the target type,
The area identification device further
An area correction unit that corrects the crossing area based on detection information about an object other than the object of the target type is provided.

The area identification device further
For a specific type of object, a position correction unit is provided that shifts the position of the object outward by a correction distance.

The area identification device further
A direction identifying unit that identifies a moving direction at each position of the crossing area is provided for each of the types.

The region identification method according to the present disclosure is
A computer detects an object of the target type from the image data obtained by the imaging device, with each of two or more types as the target type,
A computer identifies an encapsulating region that encloses the position of an object for the detected type of target;
A computer determines an overlapping region in which the connotative region specified for one type of the two or more types and the connotative region specified for another type overlap, and an area in which the object of the certain type moves. It is specified as a traversing area traversed by objects of other types.

The area identification program according to the present disclosure is
An object detection process for detecting an object of the target type from image data obtained by an imaging device, with each of two or more types as a target type;
an enclosing area specifying process for specifying an enclosing area enclosing the position of the object for the type of the object detected by the object detecting process;
The object of the certain type moves in a superimposed region where the connotative region identified for one type among the two or more types and the connotative region identified for the other type overlap by the connotative region identification processing. The computer is caused to function as an area identification device that performs overlapping area identification processing for identifying the area as a crossing area traversed by the object of the other type.

In the present disclosure, mainly moving types of objects and traversing types of objects are detected, and a traversing area is identified from the detected positions. This makes it possible to identify the crossing area in the image data without manual intervention.

1 is a configuration diagram of an area identification device 10 according to Embodiment 1. FIG. 4 is a flowchart showing the overall operation of the region identification device 10 according to Embodiment 1; 4 is a flowchart of object detection processing according to the first embodiment; FIG. 4 is an explanatory diagram of the position of a detection frame according to Embodiment 1; FIG. 4 is an explanatory diagram of the position of a detection frame according to Embodiment 1; 4 is an explanatory diagram of a two-dimensional coordinate space 42 according to Embodiment 1. FIG. 4 is a flowchart of region identification processing according to the first embodiment; FIG. 5 is an explanatory diagram of a superimposed region 52 according to Embodiment 1; FIG. 2 is a configuration diagram of an area identification device 10 according to Modification 1; FIG. 11 is a configuration diagram of an area identification device 10 according to Modification 3; FIG. 2 is a configuration diagram of an area identification device 10 according to Embodiment 2; 6 is a flowchart showing the overall operation of the area specifying device 10 according to Embodiment 2; FIG. 2 is a configuration diagram of an area identification device 10 according to Embodiment 3; 11 is a flow chart of region identification processing according to the third embodiment; FIG. 11 is an explanatory diagram of area correction processing according to the third embodiment; FIG. 11 is an explanatory diagram of area correction processing according to the third embodiment; FIG. 11 is a configuration diagram of an area identification device 10 according to Modification 5; 14 is a flowchart of region identification processing according to modification 5; FIG. 11 is an explanatory diagram of position correction processing according to Modification 5; FIG. 10 is a configuration diagram of an area specifying device 10 according to Embodiment 4; 11 is a flow chart showing the overall operation of the area specifying device 10 according to the fourth embodiment;

Embodiment 1.
*** Configuration description ***
A configuration of an area identification device 10 according to Embodiment 1 will be described with reference to FIG.
The area identifying device 10 is a computer for identifying a crossing area 53 where another object traverses a moving area of a certain type of object.
The area identification device 10 includes hardware including a processor 11 , a memory 12 , a storage 13 and a communication interface 14 . The processor 11 is connected to other hardware via signal lines and controls these other hardware.

The processor 11 is an IC (Integrated Circuit) that performs processing. Specific examples of the processor 11 are a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and a GPU (Graphics Processing Unit).

The memory 12 is a storage device that temporarily stores data. Specific examples of the memory 12 include SRAM (Static Random Access Memory) and DRAM (Dynamic Random Access Memory).

The storage 13 is a storage device that stores data. A specific example of the storage 13 is an HDD (Hard Disk Drive). In addition, the storage 13 is SD (registered trademark, Secure Digital) memory card, CF (Compact Flash, registered trademark), NAND flash, flexible disk, optical disk, compact disk, Blu-ray (registered trademark) disk, DVD (Digital Versatile Disk), etc. It may be a portable recording medium.

The communication interface 14 is an interface for communicating with an external device. Specific examples of the communication interface 14 are Ethernet (registered trademark), USB (Universal Serial Bus), and HDMI (registered trademark, High-Definition Multimedia Interface) ports.

The area identification device 10 is connected to the imaging device 31 via the communication interface 14 . The photographing device 31 is installed so as to include at least part of the crossing area to be monitored in the photographing area.

The area identification device 10 includes an initial setting unit 21, an image acquisition unit 22, an object detection unit 23, and an area identification unit 24 as functional components. The area identifying section 24 includes an enclosing area identifying section 241 and an overlapping area identifying section 242 . The function of each functional component of the area specifying device 10 is realized by software.
The storage 13 stores a program that implements the function of each functional component of the area identification device 10 . This program is read into the memory 12 by the processor 11 and executed by the processor 11 . Thereby, the function of each functional component of the area identification device 10 is realized.

Only one processor 11 is shown in FIG. However, there may be a plurality of processors 11, and the plurality of processors 11 may cooperate to execute programs that implement each function.

***Description of operation***
The operation of the area identification device 10 according to the first embodiment will be described with reference to FIGS. 2 to 8. FIG.
The operation procedure of the area identifying device 10 according to the first embodiment corresponds to the area identifying method according to the first embodiment. Also, a program that realizes the operation of the area identifying device 10 according to the first embodiment corresponds to the area identifying program according to the first embodiment.

The overall operation of the area identification device 10 according to Embodiment 1 will be described with reference to FIG.
(Step S11: Initial setting process)
The initial setting unit 21 allows the user to specify the type of object to be detected. Specifically, the initial setting unit 21 designates the type of one or more objects that mainly move and the type of one or more objects that traverse the region in which the mainly moving objects move. For example, when a no-entry area is specified as the crossing area 53, a person or a vehicle is specified as the type of object that mainly moves, and an object other than an object that mainly moves is specified as the type of object to cross. When specifying a pedestrian crossing area as the crossing area 53, a vehicle is specified as the type of the mainly moving object, and a person is specified as the type of the crossing object. As the vehicle, more specific types such as a four-wheeled vehicle, a two-wheeled vehicle, and a bicycle may be specified.
The initial setting unit 21 sets the specified type list 41 and the two-dimensional coordinate space 42 for each specified type in the memory 12 . The two-dimensional coordinate space 42 may be in any format as long as it can manage coordinate positions, and may be data in CSV (Comma-Separated Values) format, for example.

(Step S12: Image Acquisition Processing)
The video acquisition unit 22 acquires video data obtained by the imaging device 31 . At this time, the image acquisition unit 22 acquires image data having an appropriate length, which is obtained during a time period other than a time period in which the imaging device 31 is exposed to direct sunlight. The appropriate length differs depending on conditions such as the type included in the list 41 set in step S11 and the location where the photographing device 31 is installed. For example, when the type is people and vehicles, and the photographing device 31 is installed in a place where people and vehicles frequently pass, the appropriate length is about 5 hours. The moderately long video data may be continuous video data, or may be intermittent video data instead of continuous video data. In any case, the video data having a moderate length may be video data or a plurality of video data having a "moderate length of total video time".
The video acquisition unit 22 decomposes the acquired video data into image data for each frame, and writes the image data into the memory 12 .

The process of step S13 is performed on the image data of each frame obtained in step S12.

(Step S13: Object detection processing)
The object detection unit 23 detects an object of the target type from the target image data using each of the two or more types included in the list 41 set in step S11 as the target type.

(Step S14: Area specifying process)
The region specifying unit 24 sets each of the two or more types included in the list 41 set in step S11 as the target type, and determines the inclusive region 51 that includes the position of the object for the target type detected in step S13. Identify. Then, the region identifying unit 24 determines whether an object of a certain type moves in an overlapping region 52 where the connotative region 51 identified for a certain type among two or more types and the connotative region 51 identified for another type overlap each other. A crossing region 53 crossed by another type of object is specified as a crossing region 53 .

The object detection process (step S13 in FIG. 2) according to the first embodiment will be described with reference to FIG.
(Step S21: Spatial reading process)
The object detection unit 23 reads the two-dimensional coordinate space 42 for each type set in step S11.

(Step S22: detection processing)
The object detection unit 23 detects an object of the target type from the target image data using each type as the target type. Specifically, the object detection unit 23 detects an object of the target type using an object detection model for detecting the target type from the image data. Thereby, the object detection unit 23 specifies a detection frame surrounding the area where the object is detected. At this time, information on the type of the detected object is associated with the detection frame. Here, it is assumed that the detection frame is rectangular.

The process of step S23 is executed for each detection frame specified in step S22.

(Step S23: plot processing)
The object detection unit 23 identifies the type corresponding to the target detection frame by referring to the information on the type associated with the target detection frame. The object detection unit 23 plots points at the positions of the target detection frames in the two-dimensional coordinate space 42 for the specified type.
The position of the detection frame is preset by the user for each type. As a specific example, as shown in FIG. 4, when the type is human, the position of the detection frame is the position of the center point of the lower side of the detection frame. As shown in FIG. 5, when the type is a vehicle, the front or back of the vehicle is detected, and the positions of the detection frame are the positions of the endpoints on the lower side of the detection frame. Among the vehicles, the position of the center point of the lower side may be plotted for passenger cars, and the position of both end points of the lower side may be plotted for large vehicles such as buses and trucks.

(Step S24: Spatial storage processing)
The object detection unit 23 writes in the memory 12 the two-dimensional coordinate space 42 for each type whose positions are plotted in step S23. As a result, as shown in FIG. 6, a two-dimensional coordinate space 42 in which points are plotted at positions where objects are detected is generated for each type.

The area specifying process (step S14 in FIG. 2) according to the first embodiment will be described with reference to FIG.
(Step S31: Spatial reading process)
The connotative region specifying unit 241 reads the two-dimensional coordinate space 42 for each type stored in step S24 of FIG.

The processes from step S32 to step S33 are executed for each of the two or more types included in the list 41 .

(Step S32: outlier removal processing)
The connotative region identification unit 241 removes outliers from the points plotted in the two-dimensional coordinate space 42 for the target type. As a specific example, the enclosing region identification unit 241 removes outliers using a method such as the k-nearest neighbor method. In addition, the connotative region identifying unit 241 removes the points of the divided spaces obtained by dividing the two-dimensional coordinate space 42 into each reference size, and the points of the divided spaces whose point density is lower than the threshold value. may be removed.

(Step S33: Connotative area specifying process)
The encapsulating region specifying unit 241 specifies the enclosing region 51 that encloses the positions of the remaining points after removing the outliers in step S32. Specifically, the connotative region identifying unit 241 identifies the connotative region 51 by executing an algorithm such as QuickHull with the target two-dimensional coordinate space 42 from which outliers have been removed as input.

(Step S34: Overlapping area specifying process)
The superimposed region identifying unit 242 determines the superimposed region 52 where the connotative region 51 identified for a certain type in step S33 and the connotative region 51 identified for another type in step S33 overlap as a region where an object of a certain type moves. is identified as the traversing region 53 traversed by other types of objects. Here, a certain type is a type of an object that mainly moves, and another type is a type of an object that traverses an area in which the mainly moving object moves.
For example, when a pedestrian crossing area is specified as the crossing area 53, the type of the mainly moving object is a vehicle, and the type of the crossing object is at least one of a person and a bicycle. Assuming that the types of crossing objects are both people and bicycles, as shown in FIG. A crossing region 53 is identified as the overlapping region 52 . Crosswalks shall include bicycle crossings.

(Step S35: Area storage processing)
The superimposed area identifying unit 242 writes area information indicating the crossing area 53 identified in step S34 to the memory 12 .

*** Effect of Embodiment 1 ***
As described above, the region identifying apparatus 10 according to Embodiment 1 mainly detects objects of a moving type and objects of a crossing type, and identifies the crossing region 53 from the detected positions. This makes it possible to identify the crossing area in the image data without manual intervention.

Further, the region identification device 10 according to Embodiment 1 identifies the crossing region 53 after removing outliers. This makes it possible to reduce the influence of erroneous detection.

***Other Configurations***
<Modification 1>
In Embodiment 1, the video acquisition unit 22 acquires video data obtained by the imaging device 31 from the imaging device 31 . However, as shown in FIG. 9, the image data obtained by the photographing device 31 may be temporarily stored in the external storage 32 and the image acquisition unit 22 may read the image data from the external storage 32 .

<Modification 2>
In Embodiment 1, the effect of erroneous detection is reduced by removing outliers using a method such as the k-nearest neighbor method. Alternatively, or in addition, the effects of false detections may be reduced by removing the positions of objects that could not be properly tracked.
Specifically, the object detection unit 23 performs tracking processing on the detected object. Tracking processing can be realized by using existing technology. Then, the enclosing region specifying unit 241 removes points indicating the positions of objects that could not be properly tracked in the tracking process. Failure to perform appropriate tracking means that the tracking is interrupted or that the tracking results in moving to a position separated by a reference distance or more between frames.

<Modification 3>
In Embodiment 1, each functional component is realized by software. However, as a modification 3, each functional component may be realized by hardware. Regarding this modification 3, the points different from the first embodiment will be described.

The configuration of the region identification device 10 according to Modification 3 will be described with reference to FIG. 10 .
When each functional component is implemented by hardware, the area identifying device 10 includes an electronic circuit 15 in place of the processor 11, memory 12 and storage 13. FIG. The electronic circuit 15 is a dedicated circuit that realizes the functions of each functional component, memory 12 and storage 13 .

The electronic circuit 15 includes a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, a logic IC, a GA (Gate Array), an ASIC (Application Specific Integrated Circuit), and an FPGA (Field-Programmable Gate Array). is assumed.
Each functional component may be implemented by one electronic circuit 15, or each functional component may be implemented by being distributed among a plurality of electronic circuits 15. FIG.

<Modification 4>
As a modification 4, some functional components may be implemented by hardware, and other functional components may be implemented by software.

The processor 11, the memory 12, the storage 13 and the electronic circuit 15 are called a processing circuit. That is, the function of each functional component is realized by the processing circuit.

Embodiment 2.
The second embodiment differs from the first embodiment in that the crossing area 53 is monitored. In the second embodiment, this different point will be explained, and the explanation of the same point will be omitted.

*** Configuration description ***
The configuration of the area identification device 10 according to the second embodiment will be described with reference to FIG.
The region identifying device 10 differs from the region identifying device 10 shown in FIG. 1 in that it includes an object detection unit 25 as a functional component. The functions of the object detection unit 25 are implemented by software or hardware, like other functional components.

***Description of operation***
The overall operation of the area identification device 10 according to the second embodiment will be described with reference to FIG.
The processing from step S11A to step S14A is the same as the processing from step S11 to step S14 in FIG.

(Step S15A: Target detection processing)
The object detection unit 25 detects a detection object from the crossing area 53 specified in step S14A in the image data acquired by the photographing device 31 . For example, it is conceivable to detect an object placed in the crossing area 53 for a reference time or longer as a detection target. Further, when a pedestrian crossing area is specified as the crossing area 53, it is conceivable that a person in the crossing area 53 is detected as a detection target even though the signal for the pedestrian crossing is red. Whether or not a signal is a red signal may be determined by detecting it from image data, or may be determined by acquiring information indicating the state of the signal from a device that controls the signal.

*** Effect of Embodiment 2 ***
As described above, the area identification device 10 according to the second embodiment monitors the crossing area 53 by detecting the detection target from the identified crossing area 53 . Since the detection target is detected after narrowing down the area to be monitored, it is possible to suppress erroneous detection and improve detection performance.

Embodiment 3.
Embodiment 3 differs from Embodiments 1 and 2 in that the crossing area 53 is corrected based on detection information about the periphery. In the third embodiment, this different point will be explained, and the explanation of the same point will be omitted.
In the third embodiment, a case in which functions are added to the first embodiment will be described. However, it is also possible to add functions to the second embodiment.

*** Configuration description ***
The configuration of the area identification device 10 according to the third embodiment will be described with reference to FIG.
The area identification device 10 differs from the area identification device 10 shown in FIG. 1 in that it includes an area correction unit 243 as a functional component. The function of the area correction unit 243 is realized by software or hardware like other functional components.

***Description of operation***
The region specifying process (step S14 in FIG. 2) according to the third embodiment will be described with reference to FIG.
The processing from step S31A to step S34A is the same as the processing from step S31 to step S34 in FIG. The processing of step S36A is the same as the processing of step S35 in FIG.

(Step S35A: area correction processing)
The area correction unit 243 corrects the crossing area 53 specified in step S34A based on the peripheral detection information. Peripheral detection information is information about an object other than the target type, which is detected from the vicinity of the object detected in step S13. Here, in step S13, objects other than the target type are also detected.

When specifying the no-entry area as the crossing area, the area correction unit 243 corrects the crossing area 53 using the position of the gate, which is the entrance/exit of the no-entry area, as the surrounding detection information. In this case, the object detection unit 23 detects the gate in step S13 of FIG. 2, and the area correction unit 243 corrects the crossing area 53 using the detected gate position. There may be some distance between the gate and the area through which primarily moving objects pass. In this case, when specifying the crossing region 53 based on the enclosing region 51 that includes the position of the object, there will be a region that is not specified as the crossing region 53 inside the gate. Therefore, the region correction unit 243 corrects the crossing region 53 so as to slightly widen toward the gate position. Thereby, it is possible to prevent the formation of a region not specified as the crossing region 53 inside the gate.

When specifying a pedestrian crossing area as the crossing area, the area correction unit 243 corrects the crossing area 53 using the position of the display on the road surface as the peripheral detection information. Specifically, the display on the road surface is a white line (roadside line) that indicates the boundary between the sidewalk and the roadway. In this case, in step S13 of FIG. 2, the object detection unit 23 detects the white line drawn on the road surface, and the area correction unit 243 corrects the crossing area 53 using the position of the detected white line.
As shown in FIG. 16, there is some distance between the area where the vehicle passes and the sidewalk. Therefore, when the crossing area 53 is specified based on the inclusive area 51 that includes the position of the vehicle, an area that is not specified as the crossing area 53 is created between the sidewalk and the crossing area 53 . Therefore, the region correction unit 243 corrects the crossing region 53 so as to slightly expand toward the position of the white line indicating the boundary with the sidewalk. As a result, it is possible to prevent an area that is not specified as the crossing area 53 from being created between the sidewalk and the sidewalk.

*** Effect of Embodiment 3 ***
As described above, the region identification device 10 according to the third embodiment corrects the crossing region 53 identified by detecting mainly moving type objects and crossing type objects based on gates, white lines, or the like. do. Thereby, the crossing area 53 can be identified with higher accuracy.

***Other Configurations***
<Modification 5>
In Embodiment 3, the crossing area 53 is corrected. However, the position of the object detected in step S13 of FIG. 2 may be corrected.

The configuration of the area identification device 10 according to Modification 5 will be described with reference to FIG. 17 .
The area identification device 10 includes a position corrector 244 as a functional component. The function of the position corrector 244 is realized by software or hardware, like other functional components.

The area specifying process (step S14 in FIG. 2) according to Modification 5 will be described with reference to FIG.
The processing of step S31B is the same as the processing of step S31 in FIG. The processing from step S33B to step S36B is the same as the processing from step S32 to step S35 in FIG.

(Step S32B: position correction processing)
The position corrector 244 corrects the position of the object detected in step S13 of FIG.
When specifying an intrusion prohibited area as a crossing area, as shown in FIG. 19, as the position of the object, the positions of both end points on the lower side of the detection frame are used. In this case, the position correction unit 244 corrects the position of the object to a position obtained by shifting the two end points of the lower side of the detection frame outward by the correction distance.
As described with reference to FIG. 15, there is some distance between the position through which the object passes and the gate. By correcting the position of the object to the position shifted outward, the position of the object becomes closer to the gate. Thereby, it is possible to prevent the formation of a region not specified as the crossing region 53 inside the gate.

Embodiment 4.
Embodiment 4 differs from Embodiments 1 to 3 in that the moving direction of an object is specified. In the fourth embodiment, this different point will be explained, and the explanation of the same point will be omitted.
In a fourth embodiment, a case in which functions are added to the first embodiment will be described. However, it is also possible to add functions to the second embodiment.

*** Configuration description ***
The configuration of the area identification device 10 according to the fourth embodiment will be described with reference to FIG.
The area identifying device 10 differs from the area identifying device 10 shown in FIG. 1 in that it includes a direction identifying section 26 as a functional component. The function of the direction identification unit 26 is realized by software or hardware, like other functional components.

***Description of operation***
The overall operation of the area identification device 10 according to the fourth embodiment will be described with reference to FIG.
The processing from step S11B to step S14B is the same as the processing from step S11 to step S14 in FIG.

(Step S15B: direction specifying process)
The direction specifying unit 26 specifies the moving direction of the object at each position of the crossing area 53 specified in step S14 for each type.
Specifically, in step S13B, it is assumed that the object detection unit 23 is performing object tracking processing. The direction identification unit 26 identifies the movement direction at each position for each type by referring to the result of the tracking process. For example, a vehicle may have a moving direction determined by lanes. In this case, the moving direction of the vehicle can be specified for each position.
Some objects may move in different directions even if they are of the same type. The direction identifying unit 26 may remove objects moving in such exceptional directions and identify the direction in which the majority of objects move. Also, depending on the type, the direction of movement may not be fixed. If the moving direction is not fixed, the direction specifying unit 26 may specify that the moving direction is not fixed.

Also, by providing the direction specifying unit 26, it is possible to specify the type of the object and the moving direction of the object. Depending on the combination of the type of the object and its moving direction, even if the object is detected as the same object, it can be recognized as an object of a different type and the process can proceed. For example, a vehicle that moves from top to bottom on the image data (eg, vehicle type 1), a vehicle that moves from right to left on the image data (eg, vehicle type 2), and a vehicle that moves from left to right (eg, vehicle It is possible for the object detection unit 23 to operate so as to detect types 3) as objects of different types.
The modification will be described with reference to the flowchart of FIG. To detect.

*** Effect of Embodiment 4 ***
As described above, the region identification device 10 according to Embodiment 4 identifies the movement direction of an object at each position in the crossing region 53 for each type. As a result, when detecting a detection target as described in the second embodiment, it is possible to detect an object moving in a direction different from the specified movement direction.

Note that "unit" in the above description may be read as "circuit", "process", "procedure", "process", or "processing circuit".

The embodiments and modifications of the present disclosure have been described above. Some of these embodiments and modifications may be combined and implemented. Also, any one or some may be partially implemented. It should be noted that the present disclosure is not limited to the above embodiments and modifications, and various modifications are possible as necessary.

10 area identification device 11 processor 12 memory 13 storage 14 communication interface 15 electronic circuit 21 initial setting unit 22 image acquisition unit 23 object detection unit 24 area identification unit 241 connotation area identification unit 242 superimposition Area identification unit 243 Area correction unit 244 Position correction unit 25 Object detection unit 26 Direction identification unit 31 Photographing device 32 External storage 41 List 42 Two-dimensional coordinate space 51 Inclusive area 52 Superimposed area 53 transversal area.

Claims (10)

an object detection unit that detects an object of the target type from image data obtained by a photographing device, with each of two or more types as a target type;
an enclosing area specifying unit that specifies an enclosing area that encloses the position of the object for the type of the target detected by the object detecting unit;
The object of the certain type moves in a superimposed region where the connotative region identified for one type among the two or more types and the connotative region identified for the other type overlap by the connotative region identifying unit. and a superimposed area identifying unit that identifies the area as a crossing area traversed by the object of the other type. 2. The enclosing area specifying unit according to claim 1, wherein the enclosing area specifying unit specifies the enclosing area including a position remaining after outliers are removed from the positions of the object for the type of the object detected by the object detecting unit. Area identification device. The object detection unit tracks the detected object,
3. The area identifying apparatus according to claim 1, wherein the enclosing area specifying unit specifies the enclosing area that includes the remaining positions after removing the positions of the object that could not be tracked. The area identification device further
4. The area identification device according to any one of claims 1 to 3, further comprising an object detection unit that detects a detection object from the crossing area identified by the superimposed area identification unit in the image data. After detecting the object, the object detection unit plots a point at a position set by the user for each type of object,
5. The area identification device according to any one of claims 1 to 4, wherein said connotative area identification unit identifies said connotative area based on plotted points. The object detection unit also detects an object other than the object of the target type,
The area identification device further
6. The area identification device according to any one of claims 1 to 5, further comprising an area correction unit that corrects the crossing area based on detection information about an object other than the object of the target type. The area identification device further
7. The area identification device according to any one of claims 1 to 6, further comprising a position correction unit that shifts the position of the object outward by a correction distance for a specific type of object. The area identification device further
8. The area identification device according to any one of claims 1 to 7, further comprising a direction identification unit that identifies a moving direction at each position of the crossing area for each of the types. A computer detects an object of the target type from the image data obtained by the imaging device, with each of two or more types as the target type,
A computer identifies an encapsulating region that encloses the position of an object for the detected type of target;
A computer determines an overlapping region in which the connotative region specified for one type of the two or more types and the connotative region specified for another type overlap, and an area in which the object of the certain type moves. A region identification method for identifying a crossing region crossed by objects of other types. An object detection process for detecting an object of the target type from image data obtained by an imaging device, with each of two or more types as a target type;
an enclosing area specifying process for specifying an enclosing area enclosing the position of the object for the type of the object detected by the object detecting process;
The object of the certain type moves in a superimposed region where the connotative region identified for one type among the two or more types and the connotative region identified for the other type overlap by the connotative region identification processing. A region identification program that causes a computer to function as a region identification device that performs superimposed region identification processing for identifying a region as a crossing region traversed by an object of the other type.

Images