Area intrusion detection determination method, system and electronic equipment for video monitoring

Abstract

The application discloses a region intrusion detection judgment method and system for video monitoring and electronic equipment, and relates to the technical field of image data processing.The application can improve the processing efficiency of video monitoring images by taking an existing neural network detection model as a picture detection model to determine the object classification result of each image in a to-be-detected video stream.And when the obtained object classification result contains a target classification, the object intrusion coefficient is determined, and then the intrusion result is determined based on the object intrusion coefficient, so that the accuracy of the region intrusion judgment can be improved, and the problems of region intrusion missed judgment and misjudgment in the prior art can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image data processing, in particular to a region intrusion detection judgment method and system for video monitoring and electronic equipment. BACKGROUND

[0002] Video monitoring is a common means of region intrusion detection. The means used in region intrusion judgment mainly include: 1) boundary touch principle; 2) object center entering principle; and 3) simple overlap area calculation. Due to the height and tilt angle of the monitoring camera, the object (such as a person) and the target region have an up-down positional relationship in planar imaging. Moreover, the object photographed by the camera is three-dimensional, and the video picture after shooting is two-dimensional. Whether there is intrusion is judged based on the planar picture, and due to the height and tilt angle of the camera, it is more difficult to accurately judge.

[0003] Suppose the object is a person. When the person is below the region, the person will be judged as an intrusion even if only the head touches or the upper body enters the target region (legs and feet are outside). When the person is above the region, the person has already stood in the target region (feet are in the target region), and since the center point is outside, it is not considered as an intrusion. In planar imaging, the circled region is an irregular polygon, and the object appears in a variable position. It is difficult for the traditional algorithm to accurately and efficiently judge whether the object has intruded into the target region.

[0004] Based on the above-mentioned method, due to the camera shooting angle and planar imaging, there will be misjudgment and omission when judging whether the object has intruded into the region. SUMMARY

[0005] To solve the above-mentioned problems existing in the prior art, the present application provides a region intrusion detection judgment method and system for video monitoring and electronic equipment.

[0006] To achieve the above-mentioned purpose, the present application provides the following solutions.

[0007] A region intrusion detection judgment method for video monitoring, comprising: delineating a target region.

[0008] Obtaining a to-be-detected video stream, and inputting each frame of image in the to-be-detected video stream into a picture detection model to obtain an object classification result; the picture detection model is a neural network detection model.

[0009] When the object classification result contains a target classification, determining an object intrusion coefficient.

[0010] Determining an intrusion result based on the object intrusion coefficient; the intrusion result includes that the object does not intrude into the target region and that the object intrudes into the target region.

[0011] The intrusion result is determined based on the relationship between the object intrusion coefficient and the intrusion threshold, specifically including: when the object intrusion coefficient is 0, the object in the video stream to be detected does not intrude into the target area.

[0012] When the object intrusion coefficient is 1, the object in the video stream to be detected intrudes into the target area.

[0013] When the object intrusion coefficient is greater than 0 and less than 1, the orientation relationship between the object and the target area in the video stream to be detected is determined, and the intrusion result is determined based on the orientation relationship.

[0014] Optionally, determining the intrusion result based on the directional relationship specifically includes: setting a directional intrusion threshold based on the target area; the directional intrusion threshold includes an upper intrusion threshold, a lower intrusion threshold, a left intrusion threshold, and a right intrusion threshold.

[0015] The corresponding directional intrusion threshold is selected based on the aforementioned directional relationship.

[0016] The intrusion result is determined based on the selected directional intrusion threshold and the object intrusion coefficient.

[0017] Optionally, the intrusion result is determined based on the selected directional intrusion threshold and the object intrusion coefficient, specifically including: when the object intrusion coefficient is less than or equal to the selected directional intrusion threshold, the object does not intrude into the target area.

[0018] When the object intrusion coefficient is greater than the selected directional intrusion threshold, the object intrudes into the target area.

[0019] Optionally, the object intrusion coefficient is the percentage of the area of ​​the target region intruded by the object to the total area of ​​the object.

[0020] According to specific embodiments provided by the present invention, the following technical effects are disclosed: The regional intrusion detection and determination method for video surveillance provided by the present invention improves the processing efficiency of video surveillance images by using an existing neural network detection model as an image detection model to determine the object classification result of each frame of the video stream to be detected. Furthermore, when the obtained object classification result includes a target classification, an object intrusion coefficient is determined, and then the intrusion result is determined based on the object intrusion coefficient, which improves the accuracy of regional intrusion determination and thus solves the problems of missed detection and false detection of regional intrusion in the prior art.

[0021] Furthermore, this invention provides a regional intrusion detection and determination system for video surveillance, which is applied to the aforementioned regional intrusion detection and determination method for video surveillance; the system includes: a target area delineation module, a target classification determination module, an intrusion coefficient determination module, and an intrusion result determination module.

[0022] The target region determining module is configured to determine a target region.

[0023] The target classification determining module is configured to obtain a video stream to be detected and input each frame of image in the video stream to be detected into a picture detection model to obtain an object classification result.

[0024] The intrusion coefficient determining module is configured to determine an object intrusion coefficient when the object classification result contains a target classification.

[0025] The intrusion result determining module is configured to determine an intrusion result based on the object intrusion coefficient, wherein the intrusion result includes that the object does not intrude into the target region and the object intrudes into the target region.

[0026] An electronic device includes a memory and a processor.

[0027] The memory is configured to store a computer program.

[0028] The processor is connected with the memory and is configured to call and execute the computer program to implement the region intrusion detection and judgment method for video monitoring provided above.

[0029] Optionally, the memory is a computer readable storage medium.

[0030] The technical effects achieved by the two implementation structures provided above are the same as the technical effects achieved by the region intrusion detection and judgment method for video monitoring provided by the present application, and thus will not be described here. BRIEF DESCRIPTION OF DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0032] Figure 1 The flowchart of the region intrusion detection and judgment method for video monitoring provided by the present application.

[0033] Figure 2 The object and region direction calculation schematic diagram provided by the present application.

[0034] Figure 3 The object and target region local perspective direction calculation schematic diagram provided by the present application. DETAILED DESCRIPTION

[0035] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort should fall into the protection scope of the present application.

[0036] The present application aims to provide a region intrusion detection judgment method, system and electronic equipment for video monitoring, which can improve the accuracy of region intrusion judgment, and further solve the problems of region intrusion missed judgment and misjudgment in the prior art.

[0037] In order to make the above objectives, characteristics and advantages of the present application more apparent, the present application will be described in further detail below with reference to the accompanying drawings and specific embodiments.

[0038] As shown in Figure 1 The region intrusion detection judgment method for video monitoring provided by the present application is implemented through the following steps.

[0039] Step 100: demarcate a target region.

[0040] Step 101: acquire a to-be-detected video stream, and input each frame of image in the to-be-detected video stream into a picture detection model to obtain an object classification result. In this step, the existing neural network detection model is mainly used to enclose the object bounding box and label the object classification on each frame of image in the to-be-detected video stream. Commonly used neural network detection models include DNN network, residual network, GoogLeNet network, yolo network and the like.

[0041] Step 102: when the object classification result contains a target classification, determine an object intrusion coefficient. If the object classification result does not contain the target classification, no processing is performed, and the next frame of image is directly detected.

[0042] Step 103: determine an intrusion result based on the object intrusion coefficient. The intrusion result includes that the object does not intrude the target region and the object intrudes the target region. The object intrusion coefficient is the percentage of the area of the object intruding the target region to the total area of the object in the image labeled by the neural network.

[0043] For example, the object intrusion coefficient ranges from 0 to 1, and based on this, when the object intrusion coefficient is 0, the object in the to-be-detected video stream does not intrude the target region.

[0044] When the object intrusion coefficient is 1, the object in the to-be-detected video stream intrudes the target region.

[0045] When the object invasion coefficient is greater than 0 and less than 1, the positional relationship between the object and the target region in the video stream to be detected is determined, and the invasion result is determined based on the positional relationship.

[0046] Further, the invasion result is determined based on the positional relationship, specifically including: setting a positional invasion threshold based on the target region. The positional invasion threshold includes an upper invasion threshold, a lower invasion threshold, a left invasion threshold, and a right invasion threshold. Among them, the four positional invasion thresholds of up, down, left and right can be set according to actual detection needs.

[0047] The corresponding positional invasion threshold is selected based on the positional relationship.

[0048] The invasion result is determined based on the selected positional invasion threshold and the object invasion coefficient. Among them, when the object invasion coefficient is less than or equal to the selected positional invasion threshold, the object does not invade the target region. When the object invasion coefficient is greater than the selected positional invasion threshold, the object invades the target region.

[0049] Further, the up, down and left and right of the object and the target region need to be calculated separately. When calculating the position of the object and the target region, an auxiliary line can be used. Taking the up and down position calculation as an example, an auxiliary line is drawn for the object in the y-axis direction, the intersection region of the auxiliary line of the object and the target region is obtained after intersection, and the coordinates of the intersection region and the original position of the object in the y-axis direction are compared. In order to more simply and efficiently determine the positional relationship between the object and the target region in the video stream to be detected, in the region invasion calculation of the present application, the invasion coefficient is used for preliminary screening, and only the objects with the coefficient greater than 0 and less than 1 are subjected to positional calculation. At this time, the object is just on the boundary of the target region, that is Figure 2 The case of object 3, at this time, object 3 has overlapped with the target region, and there is no need to draw an extension line.

[0050] When the object is located on the boundary of the target region, it will be naturally divided into two parts: the invaded part and the non-invaded part, wherein the invaded part is the overlapping part. The positional relationship between the object and the target region is the positional relationship between the invaded part and the object. Whether it is an object or an object invasion part, they are all planar figures. In order to calculate conveniently, the centroid of the figure is used to replace the object for calculation.

[0051] The object and the target region are both polygons in planar imaging, and the object A and the target region B are regarded as polygon A and polygon B. When polygon A and polygon B intersect, the part that invades polygon B is called polygon C, and the part that does not enter is called polygon D. As Figure 3 shown, C is the invaded part of object A, D is the non-invaded part, the centroid of C is located in C, the centroid of D is located in D, and the centroid of object A is located between the centroid of C and the centroid of D.

[0052] The positional relationship between the centroid of C and the centroid of object A is equivalent to the positional relationship between the centroid of C and the centroid of D. Since the area of D is smaller than that of object A, the centroid of D is calculated faster, and D is used to replace A. At this time, the positional relationship between the local area of polygon A and polygon B is converted into the positional relationship between polygon C and polygon D. At this time, the object intrusion coefficient is the ratio of the area of C to the area of (C+D), that is, the percentage of the intrusion area in the total area of the object.

[0053] Further, the present application provides a region intrusion detection determination system for video monitoring, which is applied to the region intrusion detection determination method for video monitoring provided above. The system comprises a target region delineation module, a target classification determination module, an intrusion coefficient determination module and an intrusion result determination module.

[0054] The target region delineation module is configured to delineate a target region.

[0055] The target classification determination module is configured to obtain a to-be-detected video stream, and input each frame of image in the to-be-detected video stream into a picture detection model to obtain an object classification result. The picture detection model is a neural network detection model.

[0056] The intrusion coefficient determination module is configured to determine an object intrusion coefficient when the object classification result contains a target classification.

[0057] The intrusion result determination module is configured to determine an intrusion result based on the object intrusion coefficient. The intrusion result comprises that an object does not intrude into a target region and an object intrudes into a target region.

[0058] Further, the present application further provides an electronic device, which comprises a memory and a processor.

[0059] The memory is configured to store a computer program.

[0060] The processor is connected with the memory, and is configured to call and execute the computer program to implement the region intrusion detection determination method for video monitoring provided above.

[0061] In addition, the computer program in the storage described above is realized in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0062] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the system disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the relevant part can be referred to the method part.

[0063] The principles and implementation manners of the present application are described by applying specific examples in this paper. The above embodiment description is only used to help understand the method of the present application and its core idea. Meanwhile, for the general technical personnel in the field, according to the idea of the present application, the specific implementation manner and application range will be changed. In summary, the content of the specification should not be understood as the limitation of the present application.

Claims

1. A method for area intrusion detection decision for video surveillance, characterized in that, include: Define the target area; The video stream to be detected is acquired, and each frame of the video stream is input into the image detection model to obtain the object classification result; The image detection model is a neural network detection model; When the object classification result contains the target classification, the object intrusion coefficient is determined; The object intrusion coefficient is the percentage of the area of ​​the target region invaded by the object to the total area of ​​the object; The intrusion result is determined based on the object intrusion coefficient; the intrusion result includes whether the object does not intrude into the target area and whether the object intrudes into the target area. The intrusion result is determined based on the relationship between the object intrusion coefficient and the intrusion threshold, specifically including: When the object intrusion coefficient is 0, the object in the video stream to be detected does not intrude into the target area; When the object intrusion coefficient is 1, the object in the video stream to be detected intrudes into the target area. When the object intrusion coefficient is greater than 0 and less than 1, the orientational relationship between the object and the target region in the video stream to be detected is determined, and the intrusion result is determined based on the orientational relationship; wherein, determining the intrusion result based on the orientational relationship specifically includes: A directional intrusion threshold is set based on the target area; the directional intrusion threshold includes an upper intrusion threshold, a lower intrusion threshold, a left intrusion threshold, and a right intrusion threshold; Select the corresponding directional intrusion threshold based on the aforementioned directional relationship; The intrusion result is determined based on the selected directional intrusion threshold and the object intrusion coefficient.

2. The method of claim 1, wherein, The intrusion result is determined based on the selected directional intrusion threshold and the object intrusion coefficient, specifically including: When the object intrusion coefficient is less than or equal to the selected directional intrusion threshold, the object does not intrude into the target area; When the object intrusion coefficient is greater than the selected directional intrusion threshold, the object intrudes into the target area.

3. A regional intrusion detection decision system for video surveillance, characterized by, The system is applied to the area intrusion detection and determination method for video surveillance as described in any one of claims 1-2; the system includes: The target area delineation module is used to delineate the target area; The target classification determination module is used to acquire the video stream to be detected and input each frame of the video stream into the image detection model to obtain the object classification result; the image detection model is a neural network detection model. The intrusion coefficient determination module is used to determine the object intrusion coefficient when the object classification result contains a target classification. An intrusion result determination module is used to determine the intrusion result based on the object intrusion coefficient; the intrusion result includes whether the object does not intrude into the target area and whether the object intrudes into the target area.

4. An electronic device, comprising: include: Memory, used to store computer programs; A processor, connected to the memory, is configured to retrieve and execute the computer program to implement the area intrusion detection and determination method for video surveillance as described in any one of claims 1-2.

5. The electronic device of claim 4, wherein, The memory is a computer-readable storage medium.

Images