JP5925047B2 - Image processing apparatus and image processing method - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing method, and more particularly, to a technique suitable for use in accurately detecting the passage of an object using video from a surveillance camera.

  Conventionally, when detecting an object that has passed through a specific location in the screen using the video of the surveillance camera, the detection is performed by tracking the object detected from the video in the screen and determining the passage of the specific location. For example, a method for detecting and tracking a specific object from video data from a surveillance camera has been proposed. For example, Patent Document 1 proposes a camera that sets a detection area in an image and tracks a person area while moving the area for an automatic focusing mechanism of the camera.

JP 61-28914 A

However, the object detected from the image is subjected to processing such as integration of a plurality of tracking objects or division of one tracking object into a plurality of parts in the course of the tracking process. Tracking is not always done. For this reason, there is a case in which a detection omission occurs when tracking an object detected from an image and determining passage through a specific portion.
SUMMARY OF THE INVENTION In view of the above-described problems, an object of the present invention is to enable highly accurate passage detection even when detection object integration / division processing is performed.

The image processing apparatus according to the present invention includes a moving object detection unit that detects a moving object from a moving image, and a moving object tracking unit that performs tracking processing by associating the moving object detected from the current image with the object detected from the previous image by the moving object detection unit. means and, if the tracking object of the previous image to be tracked by the moving body tracking means is divided into a plurality in the current image, using the corresponding relationship between a plurality of tracking objects of the current image and the tracking object of the previous image And determining means for determining whether or not each of the movement trajectories of the tracking object of the previous image and the plurality of tracking objects of the current image intersects a passage detection line .
The image processing apparatus according to the present invention performs tracking processing by associating a moving object detection unit that detects a moving object from a moving image, a moving object detected from the current image by the moving object detection unit, and an object detected from the previous image. When the moving object tracking means and a plurality of tracking objects of the previous image tracked by the moving object tracking means are integrated in the current image, the correspondence between the tracking object of the previous image and the tracking object of the current image is used, And determining means for determining whether or not each of the movement trajectories of the plurality of tracking objects of the previous image and the tracking object of the current image intersects with a passage detection line .

  According to the present invention, it is possible to detect passage with high accuracy even when the tracking object is divided or integrated.

1 is a block diagram illustrating a configuration example of an image processing apparatus according to a first embodiment. It is a flowchart explaining the operation | movement procedure of an image processing apparatus. It is a figure explaining the tracking process performed in a moving body tracking part. It is a flowchart explaining the procedure of the passage detection determination performed in a state determination part. It is a figure for demonstrating the passage detection determination process of the case where a tracking object divides | segments. It is a figure for demonstrating the passage detection determination process of the case where a tracking object integrates.

Hereinafter, the present invention will be described in detail based on preferred embodiments with reference to the accompanying drawings. The configurations shown in the following embodiments are merely examples, and the present invention is not limited to the illustrated configurations.
<First Embodiment>
FIG. 1 is a block diagram illustrating a configuration example of an image processing apparatus 100 according to an embodiment of the present invention. The image processing apparatus 100 according to the present embodiment includes an image acquisition unit 101, a moving object detection unit 102, a moving object tracking unit 103, a state determination unit 104, and a parameter setting unit 105. The image processing apparatus 100 having such a configuration can be realized in a computer system including a CPU, a ROM, and a RAM by developing a program recorded in the ROM into a RAM that functions as a work memory and executing the program. it can.

The image acquisition unit 101 acquires images to be processed (images forming a moving image) from the external monitoring camera 150 in time series. An external device for acquiring an image is not limited to the monitoring camera 150, and may be an apparatus stored in a server or an external memory.
The moving object detection unit 102 detects a moving object from the image acquired by the image acquisition unit 101 by the background subtraction method. The detected moving object information includes a position on the screen, a circumscribed rectangle, and an object size. The moving object detection unit 102 is a process for detecting a motion of an object or the like from an image, and is not limited to the background subtraction method, and may be a process using a motion vector or the like. In addition, even if the detection target is stationary for a while, it may be a process for detecting an object.

The moving object tracking unit 103 performs the tracking process by associating the moving object detected in the previous image with the moving object detected in the current image. The newly detected moving object is newly given a unique tracking ID, and the moving object associated with the previous image is given the tracking ID given by the previous image.
The parameter setting unit 105 sets the detection line of the state determination unit 104.
The state determination unit 104 determines whether the object tracked by the moving object tracking unit 103 has passed the detection line set by the parameter setting unit 105. The object passage is determined based on whether or not the line segment constituting the detection line intersects with the trajectory of the object between the previous image and the current image.

FIG. 2 shows a flowchart for explaining the operation procedure of the image processing apparatus 100 of the present embodiment.
In step S <b> 201, the image acquisition unit 101 acquires an image from the monitoring camera 150, and the moving object detection unit 102 performs a moving object detection process.
In S202, the moving object detection unit 102 determines whether a moving object is detected. If no moving object is detected, the process ends. If a moving object is detected, the process proceeds to S203 to perform detection area creation processing. The detection area creation process creates an area for the next moving object tracking process, and outputs a moving object detection result effective for the tracking process.
In S204, moving object tracking processing is performed by the moving object tracking unit 103, and tracking ID, coordinates, size, event information in the current image, and the like are output as tracking object information.

FIG. 3 is a diagram for explaining the tracking process performed in S204.
In FIG. 3, 301 indicates the processing result of the previous image, and 302 indicates the processing result of the current image. In the processing result 301 in the previous image and the processing result 302 in the current image, the moving object detection unit 102 detects the first detection object 303 and the second detection object 304. These detected objects 303 and 304 are subjected to a moving object tracking process by the moving object tracking unit 103. However, since the first detected object 303 and the second detected object 304 are close to each other, they are regarded as one tracking object. Processed. In the processing result 301 in the previous image, reference numeral 305 denotes a tracking frame A that is tracking object information output from the moving object tracking unit 103, and the information includes tracking ID1, coordinates on the screen, and size information.

  In the image in the processing result 302 of the current image, the first detection object 303 moves to the upper right, and the second detection object 304 moves to the lower left. Since the distance between the detection objects of the first detection object 303 and the second detection object 304 is greater than a certain threshold, the moving object tracking unit 103 processes these as individual tracking objects. Reference numeral 306 denotes a tracking frame B (306) for tracking the first detection object 303, 307 denotes a tracking frame C for the second detection object 304, and the tracking frame A (305) follows the tracking frame B (306). It is divided into a plurality of frames C (307).

  At this time, a parent-child relationship of the tracking frame is generated between the tracking frame B (306) and the tracking frame C (307), and the tracking frame B (306) on the side where the tracking ID is continued is the parent tracking frame and the new tracking ID. A tracking frame C (307) to which is assigned is a child tracking frame. In the tracking object information in the tracking frame B (306), the tracking ID is ID1, and the information of “divide and continue” is held as event information. In addition, since the tracking frame C (307) is newly created by dividing the tracking frame A (305), the tracking ID is newly assigned ID2 as the tracking object information, and “divided as event information” And “Newly created” information is retained.

Returning to the flowchart of FIG. As described above, when the tracking processing is performed by the moving body tracking unit 103, the image processing apparatus 100 manages tracking result information in S205. The tracking object information output by the moving object tracking unit 103 is managed, and the tracking object information history, tracking object information array, tracking object parent-child relationship, and the like are managed.
In S206, the state determination unit 104 determines the state of the object. In the state determination, for example, it is determined whether or not the tracking object has passed through the passage detection line set by the parameter setting unit 105, and passage or intrusion is detected.

  FIG. 4 is a flowchart for explaining the procedure of passage detection determination performed by the state determination unit 104. As an example of the movement pattern of the object for which passage determination is performed, a case where the tracking object is divided as shown in FIGS. 5 and 6 and a case where the tracking object is integrated will be exemplified.

  FIG. 5 is a diagram for explaining processing for performing passage detection determination in the case where the tracking object described in FIG. 3 is divided. In FIG. 5, reference numeral 501 denotes a passage detection line set by the parameter setting unit 105. When the object passes through the passage detection line from the top to the bottom of the image, the state determination unit 104 determines that the passage is detected.

FIG. 6 is a diagram for explaining processing for performing passage detection determination in a case where two tracking objects are integrated into one.
In FIG. 6, reference numeral 601 indicates the processing result of the previous image, and reference numeral 602 indicates the processing result of the current image. Reference numeral 603 denotes a third detection object detected by the moving object detection unit 102, and reference numeral 604 denotes a fourth detection object. These are tracked by the moving body tracking unit 103, and a tracking frame D (605) and a tracking frame E (606) as tracking object information are shown. The tracking frame D (605) is assigned ID3 as the tracking ID, and the tracking ID of the tracking frame E (606) is assigned ID4.

  Reference numeral 608 denotes a passage detection line set by the parameter setting unit 105. The third detection object 603 has moved to the lower left side in the processing result 602 of the current image, and the fourth detection object 604 has moved to the right side in the current image. Therefore, since the third detection object 603 and the fourth detection object 604 approach a distance equal to or less than a certain threshold in the current image, the moving object tracking unit 103 sets the tracking frame D (605) and the tracking frame E (606) to the tracking frame. F (607) is integrated. The tracking frame F (607) continues the tracking frame E (606), the tracking ID is ID4, and holds “integrated and continued” information as event information. As for the parent-child relationship by integration, the tracking ID 4 is the parent tracking frame and the tracking ID 3 is the child tracking frame, and the information is held in the parent-child relationship management of tracking result information management.

Next, the passage detection determination operation by the state determination unit 104 will be described with reference to FIG.
The state determination unit 104 performs passage determination processing for all tracking frames in the current image according to the procedure shown in the flowchart of FIG.
In step S <b> 401, the state determination unit 104 refers to the event information of the tracking object information and determines whether the frame is a new frame without division / integration. If the frame is a new frame with no division / integration event, there is no movement trajectory of the object, so the process proceeds to S408, the passage is not detected, and the process ends.

  In S401, if it is not a tracking frame without division / integration, the process proceeds to S402 to determine whether the frame is a new frame by division. The tracking frame C (307) in FIG. 5 corresponds to this because it is a tracking frame newly created by division.

In S403, the state determination unit 104 refers to the information of the parent tracking frame before division managed by the tracking information result management, and calculates a trajectory with the tracking frame.
In S407, it is determined whether or not the calculated trajectory intersects with the line segment constituting the passage detection line. If so, the process proceeds to S409 to detect passage. If the vehicle does not intersect, the process proceeds to S408, where it is determined that the passage is not detected (S408), and the process ends.
In the tracking frame C (307) in FIG. 5, the calculated trajectory line is represented as a tracking frame trajectory B (503). Since the tracking frame trajectory B (503) intersects the passage detection line 501, in this case, it is determined as passage detection.
As described above, the state is determined using the correspondence relationship between the tracking object before the division and the plurality of tracking objects after the division.

  On the other hand, in S402, if it is not a new frame by division, it is determined in S404 whether or not an integration event has occurred. The tracking frame F (607) in FIG. 6 corresponds to the integrated tracking frame obtained by integrating the tracking frame D (605) and the tracking frame E (606).

In step S405, the state determination unit 104 calculates the disappearance frame (the tracking frame D (605) in FIG. 6 and the tracking frame trajectory in FIG. 6) managed by the tracking information result management. Next, in step S406. The track of the tracking frame between the current image and the previous image of the tracking frame is calculated.
In S407, it is determined whether or not these trajectories intersect with the line segments constituting the detection line. If they intersect, the process proceeds to S409 to detect the passage, and if not, the process proceeds to S408 and the passage is not detected. Determine and end the process.

In the tracking frame F (607) in FIG. 6, a trajectory with the tracking frame D (605), which is a lost frame by integration, is calculated (processing of S405).
Reference numeral 609 denotes a third tracking frame locus from the tracking frame D (605) to the tracking frame F (607). In addition, in order to calculate the track of the tracking frame (S406 processing), the track with the tracking frame E (606) is calculated.
Reference numeral 610 denotes a fourth tracking frame locus from the tracking frame E (606) to the tracking frame F (607). Next, these add tail frame locus determines whether intersects the detection line (S407) is a third tracking frame track 609, since the cross the detection line, and passage detection (S409) To be judged. Further, since the fourth tracking locus 610 does not intersect the detection line, it is determined that the passage detection is not detected (S408), and the state determination process ends.
Thus, the state is determined using the correspondence between the tracking object before integration and the tracking object after integration.

If it is determined in S404 that an integrated event has not occurred, the state determination unit 104 calculates the tracking frame trajectory between the previous image and the current image (S406). The tracking frame B (306) in FIG. 5 corresponds to this case. A trajectory from the tracking frame A (305) to the tracking frame B (306) is indicated by a tracking frame trajectory A (502).
Next, it is determined whether or not the calculated trajectory and the passage detection line intersect (S407), whether passage is detected or not detected (S408, S409), and the process is terminated. In the case of the tracking frame B (306), since it does not intersect with the passage detection line, it is determined that the passage is not detected, and the state determination process ends.

  Thus, when determining the passage detection of the object, even if the detection object is divided or integrated, by referring to the respective tracking object information before the division or the respective tracking object information before the integration, It is possible to eliminate leakage of passage detection, and it is possible to realize passage detection with higher accuracy.

<Other embodiments>
Although the embodiment has been described in detail above, the present invention can take an embodiment as a system, apparatus, method, program, recording medium (storage medium), or the like. Specifically, the present invention may be applied to a system composed of a plurality of devices (for example, a host computer, an interface device, an imaging device, a web application, etc.), or may be applied to a device composed of a single device. Good.

  Needless to say, the object of the present invention can be achieved as follows. That is, a recording medium (or storage medium) in which a program code (software for controlling a computer) that realizes the functions of the above-described embodiments is recorded is supplied to the system or apparatus. Needless to say, such a storage medium is a computer-readable storage medium. Then, the computer (or CPU or MPU) of the system or apparatus reads and executes the program code stored in the recording medium. In this case, the program code itself read from the recording medium realizes the functions of the above-described embodiment, and the recording medium on which the program code is recorded constitutes the present invention.

301 Processing result in previous image 302 Processing result in current image 303 First detection object 304 Second detection object 305 Tracking frame A
306 Tracking frame B
307 Tracking frame C
501 Passing detection line 502 Tracking frame locus A
503 Tracking frame trajectory B
601 Processing result in previous image 602 Processing result in current image 603 Third detection object 604 Fourth detection object 605 Tracking frame D
606 Tracking frame E
607 Tracking frame F
608 Passing detection line 609 Third tracking frame locus 610 Fourth tracking frame locus

Claims (6)

A moving object detection means for detecting a moving object from a video;
A moving body tracking means for performing tracking processing in association with a moving body detected from the current image by the moving body detection means and an object detected from the previous image ;
Wherein when the tracking object of the previous image to be tracked by the moving body tracking means is divided into a plurality in the current image, using the corresponding relationship between a plurality of tracking objects of the current image and the tracking object of the previous image, the front An image processing apparatus, comprising: a determination unit configured to determine whether each of movement traces of a tracking object of an image and a plurality of tracking objects of the current image intersects a passage detection line . A moving object detection means for detecting a moving object from a video;
A moving body tracking means for performing tracking processing in association with a moving body detected from the current image by the moving body detection means and an object detected from the previous image ;
If multiple tracking objects of the previous image to be tracked by the moving body tracking means is integrated in the current image, using the correspondence between the tracked object of the current image and the tracking object of the previous image, a plurality of said previous image An image processing apparatus comprising: a determination unit configured to determine whether each of the movement trajectories of the tracking object and the tracking object of the current image intersects a passage detection line . A moving object detection step of detecting a moving object from a video;
A moving body tracking step of performing tracking processing in association with the moving body detected from the current image and the object detected from the previous image in the moving body detection step;
When said tracking objects of the previous image to be tracked in the moving object tracking step is divided into a plurality in the current image, using the corresponding relationship between a plurality of tracking objects of the current image and the tracking object of the previous image, the front An image processing method comprising: a determination step of determining whether or not each movement locus of a tracking object of an image and a plurality of tracking objects of the current image intersects a passage detection line . A moving object detection step of detecting a moving object from a video;
A moving body tracking step of performing tracking processing in association with the moving body detected from the current image and the object detected from the previous image in the moving body detection step;
If multiple tracking objects of the previous image to be tracked in the moving object tracking step is integrated in the current image, using the correspondence between the tracked object of the current image and the tracking object of the previous image, a plurality of said previous image And a determination step of determining whether each of the movement trajectories of the tracking object of the current image and the tracking object of the current image intersects the passage detection line . A moving object detection means for detecting a moving object from a video;
A moving body tracking means for performing tracking processing in association with a moving body detected from the current image by the moving body detection means and an object detected from the previous image ;
Wherein when the tracking object of the previous image to be tracked by the moving body tracking means is divided into a plurality in the current image, using the corresponding relationship between a plurality of tracking objects of the current image and the tracking object of the previous image, the front A program for causing a computer to function as a determination unit that determines whether each of the movement trajectories of a tracking object of an image and a plurality of tracking objects of the current image intersects a passage detection line . A moving object detection means for detecting a moving object from a video;
A moving body tracking means for performing tracking processing in association with a moving body detected from the current image by the moving body detection means and an object detected from the previous image ;
If multiple tracking objects of the previous image to be tracked by the moving body tracking means is integrated in the current image, using the correspondence between the tracked object of the current image and the tracking object of the previous image, a plurality of said previous image A program for causing a computer to function as determination means for determining whether each of the movement trajectories of the tracking object and the tracking object of the current image intersects the passage detection line .

Images