Patrol behavior monitoring method and device for patrol personnel

Abstract

The invention discloses a patrol behavior monitoring method and device for patrol personnel, and relates to the technical field of patrol recognition, and the method comprises the steps: generating a unique identification ReID code of a target patrol personnel in a patrol region when the target patrol personnel is recognized to enter the patrol region; acquiring patrol information of all patrol areas within a set time, and analyzing the patrol information to obtain unique identification ReID codes of patrol personnel in each patrol area and corresponding patrol path information; target patrol path information of the target patrol personnel is obtained by matching the target patrol personnel with the unique identification ReID codes of the patrol personnel; and when the patrol paths contained in the target patrol path information all meet the preset patrol requirements, judging that the patrol behavior of the target patrol personnel is in place. According to the invention, the problem in the prior art that whether the patrol personnel actually and effectively carry out patrol work cannot be accurately identified when the patrol personnel carry out patrol is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of patrol identification, and in particular to a method and device for monitoring patrol behavior of patrol personnel. Background Art

[0002] In the prior art, when patrol officers are on patrol, in order to monitor such patrol behavior, generally corresponding patrol checkpoints are set up. For example, biometric patrol attendance punch-in points are deployed in patrol corridors. Patrol officers need to punch in at the attendance punch-in points every time they patrol, so as to determine whether the patrol officers are patrolling normally. However, in some cases, patrol officers do not actually carry out patrols, but directly punch in at the attendance punch-in points, and patrols are not truly and effectively carried out, resulting in inadequate patrols. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a method and device for monitoring the patrol behavior of patrol personnel, aiming to solve the problem in the prior art that patrol personnel cannot accurately identify whether they are truly and effectively carrying out their patrol work during patrol.

[0004] An object of the present invention is to provide a patrol behavior monitoring method for patrol personnel, which is used to monitor the patrol behavior of patrol personnel in a patrol area, and each patrol aisle in the patrol area is equipped with an image acquisition device, and the method comprises: When a target patrol officer is identified to enter the patrol area, the patrol route of the target patrol officer is tracked and recorded, and a unique identification code ReID code of the identity characteristics of the target patrol officer in the patrol area is generated through a body feature recognition algorithm; Acquire the patrol information of all patrol areas collected by the image acquisition device within a set time, and analyze the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel of each patrol area and the corresponding patrol path information; The target patrol path information of the target patrol person in all patrol areas is obtained by matching the unique identification ReID code of the identity feature of the target patrol person with the unique identification ReID code of the patrol person; When the patrol paths included in the target patrol path information all meet the preset patrol requirements, it is determined that the patrol behavior of the target patrol personnel is in place.

[0005] Furthermore, in the above-mentioned patrol behavior monitoring method of patrol personnel, the steps of tracking and recording the patrol route of the target patrol personnel and generating a unique identification ReID code of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm include: When capturing a preset road section in a patrol area, a patrol video of a target patrol person patrolling the preset road section; The patrol video is analyzed to determine the gait and appearance characteristics of the target patrol personnel, and a unique identification ReID code of the identity characteristics of the target patrol personnel is generated according to the gait and appearance characteristics of the target patrol personnel.

[0006] Furthermore, in the above-mentioned patrol behavior monitoring method of patrol personnel, wherein the patrol information is a patrol video, the steps of obtaining the patrol information of all patrol areas collected by the image acquisition device within a set time, and analyzing the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel of each patrol area and the corresponding patrol path information include: Decompose the patrol video into continuous video frames, obtain the first video frame where the patrol person appears, identify the first video frame to obtain the initial position of the patrol person, and mark the area of ​​the patrol person to obtain a target template; In each subsequent video frame of the first video frame, the target template and the trained patrol personnel recognition model are used to identify the video frame to obtain subsequent position information of the patrol personnel; The position information of the patrol personnel is continuously recorded, and the time spent at each location is obtained to generate the patrol path information of the patrol personnel.

[0007] Furthermore, in the above-mentioned patrol behavior monitoring method of patrol personnel, the training process of the patrol personnel identification model is: Collect continuous historical video frames containing patrol personnel, and mark the position of patrol personnel in each video frame in the historical video frames to obtain a target template, and form a training data set with the target template and the corresponding candidate area selected in the video frame; Design a weight-sharing twin network architecture, the two inputs of which are the template image of the target template and the candidate region of the current video frame, and the output of which is the similarity score; The twin network architecture is iteratively trained using a contrast loss function and a training data set until the network parameters tend to be stable to obtain the patrol personnel recognition model.

[0008] Furthermore, in the above-mentioned patrol behavior monitoring method of patrol personnel, the formula of the contrast loss function is: ; in, N is the number of samples, y i It's a label. y i = 1 means the two input images are similar, y i = 0 means the two input images are not similar,D is the distance between the current candidate region and the target template, m is a hyperparameter representing the minimum distance between non-similar samples. i Indicates i samples. Furthermore, in the above-mentioned patrol behavior monitoring method of patrol personnel, the target patrol path information includes the location information of each point on the patrol path of each patrol area and the corresponding arrival time and stay time, and when the patrol paths included in the target patrol path information meet the preset patrol requirements, the step of determining whether the patrol behavior of the target patrol personnel is in place includes: Using the position information of each point on the patrol path as the ordinate and the corresponding arrival time as the abscissa, the patrol path is mapped to a preset coordinate system to obtain a patrol path line; Obtain the target position and corresponding dwell time of the patrol route, determine the corresponding target point according to the dwell time based on the preset rules, and highlight the target point on the patrol route to obtain the target patrol route; The target patrol route is compared with a standard patrol route, and when the similarity between the target patrol route and the standard patrol route is higher than a threshold, it is determined that the patrol behavior of the target patrol person is in place.

[0009] Furthermore, in the above-mentioned patrol behavior monitoring method of patrol personnel, the step of comparing the target patrol path line with the standard patrol path line and determining that the patrol behavior of the target patrol personnel is in place when the similarity between the target patrol path line and the standard patrol path line is higher than a threshold includes: Intercepting a preset portion from the target patrol route line and comparing it with a corresponding portion of the standard patrol route line; After the comparison is passed, the remaining target patrol path line is compared with the standard patrol path line, and the patrol behavior of the target patrol person is determined after the remaining target patrol path line is compared with the standard patrol path line.

[0010] Another object of the present invention is to provide a patrol behavior monitoring device for patrol personnel, which is used to monitor the patrol behavior of patrol personnel in patrol areas. Each patrol aisle in each patrol area is equipped with an image acquisition device, and the device includes: The recording module is used to track and record the patrol route of the target patrol personnel when it is identified that the target patrol personnel enters the patrol area, and to generate a unique identification code ReID of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm; An acquisition module is used to acquire the patrol information of all patrol areas acquired by the image acquisition device within a set time, and analyze the patrol information of each patrol area to obtain a unique identification ReID code of the patrol personnel of each patrol area and corresponding patrol path information; A matching module is used to match the unique identification ReID code of the target patrol personnel's identity characteristics with the unique identification ReID code of the patrol personnel to obtain the target patrol path information of the target patrol personnel in all patrol areas; The monitoring module is used to determine that the patrol behavior of the target patrol personnel is in place when the patrol paths included in the target patrol path information all meet the preset patrol requirements.

[0011] Another object of the present invention is to provide a readable storage medium having a computer program stored thereon, wherein the program implements the steps of the above method when executed by a processor.

[0012] Another object of the present invention is to provide an electronic device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above method when executing the program.

[0013] The present invention tracks and records the patrol route of the target patrol personnel when the image acquisition device recognizes that the target patrol personnel enters the patrol area, and generates a unique identification ReID code of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm; obtains the patrol information of all patrol areas within a set time, and analyzes the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel in each patrol area and the corresponding patrol path information; matches the unique identification ReID code of the identity characteristics of the target patrol personnel with the unique identification ReID code of the patrol personnel, and obtains the unique identification ReID code of the target patrol personnel in all patrol areas. The target patrol path information is obtained; when the patrol paths included in the target patrol path information meet the preset patrol requirements, the patrol behavior of the target patrol personnel is determined to be in place, and the patrol information in the entire patrol process is collected by the image acquisition device, and the entire patrol information is identified to determine whether the patrol behavior is in place. It is not just to identify and judge whether the patrol behavior is in place through a punch-in data, which improves the accuracy of patrol behavior identification and monitoring, and uses the unique identification ReID code generated by the identity characteristics of the target patrol personnel to accurately extract the patrol information of each target patrol personnel. When facing multiple patrols, it can still accurately identify and judge whether the patrol behavior is in place. It solves the technical problem in the prior art that patrol personnel cannot accurately identify whether patrol personnel are truly and effectively carrying out patrol work during patrol. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1It is a flow chart of a method for monitoring patrol behavior of patrol personnel proposed in the first embodiment of the present invention; Figure 2 This is a structural block diagram of a patrol behavior monitoring device for patrol personnel proposed in the third embodiment of the present invention.

[0015] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0016] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0018] Embodiment 1 See also Figure 1 , shown is a method for monitoring patrolling behavior of patrolling personnel in a first embodiment of the present invention, which is used to monitor the patrolling behavior of patrolling personnel in a patrolling area. An image acquisition device is arranged in the patrol corridor of each patrolling area. The method includes steps S10 to S13.

[0019] Step S10, when it is identified that the target patrol person enters the patrol area, the patrol route of the target patrol person is tracked and recorded, and a unique identification ReID code of the identity characteristics of the target patrol person in the patrol area is generated through a body feature recognition algorithm.

[0020] Among them, an image acquisition device is set on the patrol aisle in each patrol area, so that the patrol process of the patrol personnel in the patrol area can be fully captured. Specifically, the image acquisition device can be a camera, and the target patrol personnel is the personnel who currently need to patrol the patrol area or the personnel whose patrol behavior needs to be monitored. A biometric device, such as fingerprint punching, can be set at the entrance of the patrol area to accurately identify whether the target patrol personnel has entered the patrol area, or the information collected by the image acquisition device at the entrance of the patrol area can be used to determine whether the target patrol personnel has entered the corresponding patrol area.

[0021] In addition, after identifying the target patrol personnel in the patrol area, the target patrol personnel are tracked and recorded, so that a unique identification ReID code representing the identity characteristics of the target patrol personnel can be generated through a body feature recognition algorithm, wherein the same person can be identified under different camera perspectives through the ReID code, so that the patrol information of the target patrol personnel in all patrol areas that need to be patrolled can be collected. Specifically, after the target patrol personnel enters the patrol area, the unique identification ReID code of the target patrol personnel can be determined when patrolling on a preset length of the beginning of the patrol area. Exemplarily, a patrol video of the target patrol personnel of the section can be obtained, and the patrol video can be analyzed and identified to obtain the gait and appearance characteristics of the target patrol personnel, wherein the appearance characteristics can be body shape, clothing, etc. In specific implementation, a gait feature recognition network, such as a convolutional neural network or a time series model, can be used to realize gait recognition, and an appearance feature extraction network can be used to realize appearance feature recognition. Finally, the recognized gait and appearance features are used to generate corresponding unique identification ReID codes representing identity characteristics.

[0022] Step S11, obtaining the patrol information of all patrol areas collected by the image acquisition device within a set time, and analyzing the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel of each patrol area and the corresponding patrol path information.

[0023] Among them, the purpose of setting the time is to limit the time for obtaining patrol information. Since the patrol behavior of patrol personnel is generally completed within a fixed time period, all patrol information generated in the patrol area is obtained within the set time. According to the patrol information, the unique identification ReID code of each patrol personnel and the corresponding patrol path information can be determined. Since in the actual patrol process, there will be different types of patrol personnel patrolling the same area, all patrol personnel and corresponding patrol path information can be obtained and stored, and then the patrol path information of the required target patrol personnel can be obtained through the matching mechanism.

[0024] Specifically, the patrol information is a patrol video collected by an image acquisition device. The patrol video is decomposed into continuous video frames, and the first video frame in which a patrol person appears is obtained. The first video frame is identified to obtain the initial position of the patrol person and the area of ​​the patrol person is marked to obtain a target template; the target template can track the target to determine the position information of the patrol person. In each subsequent video frame of the first video frame, the target template and the trained patrol person recognition model are used to identify the video frame to obtain the subsequent position information of the patrol person. Through the characteristics of the target template, the patrol person in the subsequent video frames can be found through similarity matching, and then the position information of the patrol person can be determined; the position information of the patrol person is continuously recorded, and the time spent at each position is obtained to generate the patrol path information of the patrol person.

[0025] More specifically, continuous historical video frames containing patrol personnel are collected, and the positions of patrol personnel are annotated in each video frame of the historical video frames to obtain target templates, and the target templates and corresponding candidate areas selected in the video frames are combined into a training data set; Design a weight-sharing twin network architecture, the two inputs of which are the template image of the target template and the candidate region of the current video frame, and the output of which is the similarity score; The twin network architecture is iteratively trained using a contrast loss function and a training data set until the network parameters tend to be stable to obtain the patrol personnel recognition model.

[0026] Among them, the Siamese Network is a model composed of two sub-networks with shared weights, which is usually used for metric learning tasks. The two sub-networks in the Siamese Network (usually convolutional neural networks) share the same parameters to process the target template and the candidate area image. The output of the network is a similarity score, which indicates the similarity between the target template and the current candidate area. The larger the similarity value, the more likely the candidate area is to be the patrol personnel corresponding to the target template. Specifically, a training data set is established. The target template image and the corresponding candidate area image can be obtained by manually marking the patrol personnel's location information in the obtained historical video frames. Then, the above-mentioned Siamese Network is used for training to enable it to have the ability to recognize the patrol personnel's location information and obtain the corresponding patrol personnel recognition model. Exemplary, the formula of the comparison loss function is: ; in, N is the number of samples, y i It's a label. y i = 1 means the two input images are similar,y i = 0 means the two input images are not similar, D is the distance between the current candidate region and the target template, which can be the Euclidean distance. m is a hyperparameter representing the minimum distance between non-similar samples. i Indicates i samples.

[0027] Step S12, the target patrol path information of the target patrol person in all patrol areas is obtained by matching the unique identification ReID code of the identity feature of the target patrol person with the unique identification ReID code of the patrol person.

[0028] Among them, the unique identification ReID code and patrol path information are obtained and stored in advance. At this time, the target patrol path information of the target patrol person can be obtained according to needs using the matching mechanism. The target patrol path information is used to judge whether the patrol behavior of the target patrol person is in place.

[0029] Step S13: when the patrol paths included in the target patrol path information all meet the preset patrol requirements, it is determined that the patrol behavior of the target patrol personnel is in place.

[0030] Among them, the target patrol path information includes the patrol path information of multiple patrol areas, that is, it may include multiple patrol paths. Each patrol path is judged and compared with the preset patrol standards, such as the duration of the patrol, whether the patrol path meets the requirements, etc. If all are met, it is determined that the patrol behavior of the target patrol personnel is in place.

[0031] In summary, the patrol behavior monitoring method of patrol personnel in the above-mentioned embodiment of the present invention tracks and records the patrol route of the target patrol personnel when the image acquisition device recognizes that the target patrol personnel enters the patrol area, and generates a unique identification ReID code of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm; obtains the patrol information of all patrol areas within a set time, and analyzes the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel in each patrol area and the corresponding patrol path information; matches the unique identification ReID code of the identity characteristics of the target patrol personnel with the unique identification ReID code of the patrol personnel, and obtains the target patrol personnel. The target patrol path information of the target patrol personnel in all patrol areas; when the patrol paths contained in the target patrol path information meet the preset patrol requirements, the patrol behavior of the target patrol personnel is determined to be in place, and the patrol information in the entire patrol process is collected by the image acquisition device, and the entire patrol information is identified to determine whether the patrol behavior is in place. It is not just to identify and judge whether the patrol behavior is in place through a punch-in data, which improves the accuracy of patrol behavior identification and monitoring, and uses the unique identification ReID code generated by the identity characteristics of the target patrol personnel to accurately extract the patrol information of each target patrol personnel. When facing multiple patrols, it can still accurately identify and judge whether the patrol behavior is in place. It solves the technical problem in the prior art that patrol personnel cannot accurately identify whether patrol personnel are truly and effectively carrying out patrol work during patrol.

[0032] Embodiment 2 This embodiment also proposes a patrol behavior monitoring method for patrol personnel. The patrol behavior monitoring method for patrol personnel proposed in this embodiment is different from the patrol behavior monitoring method for patrol personnel proposed in the first embodiment in that: The target patrol path information includes the location information of each point on the patrol path of each patrol area and the corresponding arrival time and stay time. When the patrol paths included in the target patrol path information meet the preset patrol requirements, the step of determining whether the patrol behavior of the target patrol personnel is in place includes: Using the position information of each point on the patrol path as the ordinate and the corresponding arrival time as the abscissa, the patrol path is mapped to a preset coordinate system to obtain a patrol path line; Obtain the target position and corresponding dwell time of the patrol route, determine the corresponding target point according to the dwell time based on the preset rules, and highlight the target point on the patrol route to obtain the target patrol route; The target patrol route is compared with a standard patrol route, and when the similarity between the target patrol route and the standard patrol route is higher than a threshold, it is determined that the patrol behavior of the target patrol person is in place.

[0033] Among them, the target patrol path information includes the patrol path location information, arrival time information and stay time of each patrol area. Specifically, the patrol path can be regarded as composed of multiple continuous location points. The specific location information of the location point is obtained as the ordinate, and the arrival time is the abscissa. The patrol path is mapped to the preset coordinate system to obtain the patrol path line. For example, the patrol aisles are generally regular straight lines. The distance information in the travel direction on the patrol path can be mapped to the ordinate as the location information, the time of the first arrived position can be set to zero, and the time interval between the subsequent and first arrived positions can be used as the respective arrival times, which are mapped to the abscissa to determine the initial patrol path line.

[0034] Furthermore, the target position for stopping and the corresponding stop time are obtained. According to the length of the stop time, the corresponding circular target point of different sizes can be obtained, and this point is mapped on the path line according to the position information to highlight it, so as to obtain the target patrol path line. According to the target patrol path line, it can be intuitively seen whether the target patrol personnel are in place for patrol. Specifically, according to the above-mentioned method for obtaining the target patrol path line, a standard patrol path line is determined through the information of the standard patrol process, and the target patrol path line is compared with the standard patrol path line. When the similarity between the target patrol path line and the standard patrol path line is higher than a threshold, it is determined that the patrol behavior of the target patrol personnel is in place. For example, the target patrol path line and the standard patrol path line are matched in terms of trend, number of target points and size, and the final result is weighted and calculated according to the matching results to determine whether they are similar.

[0035] In addition, in some optional embodiments of the present invention, when performing the comparison, a preset portion can be intercepted from the target patrol route line and compared with the corresponding portion of the standard patrol route line; after the comparison is successful, the remaining target patrol route line is compared with the standard patrol route line, and the patrol behavior of the target patrol personnel is determined after the remaining target patrol route line is compared with the standard patrol route line. In order to improve the efficiency of the comparison, one of the fragments is matched first, and the remaining matches are performed after the match is successful. If the match fails at the beginning, it can be directly confirmed that the target patrol personnel is not in place for patrol. In summary, the patrol behavior monitoring method of patrol personnel in the above-mentioned embodiment of the present invention tracks and records the patrol route of the target patrol personnel when the image acquisition device recognizes that the target patrol personnel enters the patrol area, and generates a unique identification ReID code of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm; obtains the patrol information of all patrol areas within a set time, and analyzes the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel in each patrol area and the corresponding patrol path information; matches the unique identification ReID code of the identity characteristics of the target patrol personnel with the unique identification ReID code of the patrol personnel, and obtains the target patrol personnel. The target patrol path information of the target patrol personnel in all patrol areas; when the patrol paths contained in the target patrol path information meet the preset patrol requirements, the patrol behavior of the target patrol personnel is determined to be in place, and the patrol information in the entire patrol process is collected by the image acquisition device, and the entire patrol information is identified to determine whether the patrol behavior is in place. It is not just to identify and judge whether the patrol behavior is in place through a punch-in data, which improves the accuracy of patrol behavior identification and monitoring, and uses the unique identification ReID code generated by the identity characteristics of the target patrol personnel to accurately extract the patrol information of each target patrol personnel. When facing multiple patrols, it can still accurately identify and judge whether the patrol behavior is in place. It solves the technical problem in the prior art that patrol personnel cannot accurately identify whether patrol personnel are truly and effectively carrying out patrol work during patrol.

[0036] Embodiment 3 See also Figure 2 , shown is a patrol behavior monitoring device for patrol personnel proposed in the third embodiment of the present invention, which is used to monitor the patrol behavior of patrol personnel in the patrol area. An image acquisition device is arranged in the patrol corridor of each patrol area, and the device includes: The recording module 100 is used to track and record the patrol route of the target patrol personnel and generate a unique identification code ReID of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm when the target patrol personnel enters the patrol area; The acquisition module 200 is used to acquire the patrol information of all patrol areas acquired by the image acquisition device within a set time, and analyze the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel of each patrol area and the corresponding patrol path information; The matching module 300 is used to match the unique identification ReID code of the target patrol personnel's identity characteristics with the unique identification ReID code of the patrol personnel to obtain the target patrol path information of the target patrol personnel in all patrol areas; The monitoring module 400 is used to determine that the patrol behavior of the target patrol personnel is in place when the patrol paths included in the target patrol path information all meet the preset patrol requirements.

[0037] Furthermore, in the patrol behavior monitoring device of the patrol personnel, the steps of tracking and recording the patrol route of the target patrol personnel and generating a unique identification ReID code of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm include: When capturing a preset road section in a patrol area, a patrol video of a target patrol person patrolling the preset road section; The patrol video is analyzed to determine the gait and appearance characteristics of the target patrol personnel, and a unique identification ReID code of the identity characteristics of the target patrol personnel is generated according to the gait and appearance characteristics of the target patrol personnel.

[0038] Furthermore, in the patrol behavior monitoring device of the patrol personnel, wherein the patrol information is a patrol video, the steps of obtaining the patrol information of all patrol areas collected by the image acquisition device within a set time, and analyzing the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel of each patrol area and the corresponding patrol path information include: Decompose the patrol video into continuous video frames, obtain the first video frame where the patrol person appears, identify the first video frame to obtain the initial position of the patrol person, and mark the area of ​​the patrol person to obtain a target template; In each subsequent video frame of the first video frame, the target template and the trained patrol personnel recognition model are used to identify the video frame to obtain subsequent position information of the patrol personnel; The position information of the patrol personnel is continuously recorded, and the time spent at each location is obtained to generate the patrol path information of the patrol personnel.

[0039] Furthermore, in the above-mentioned patrol behavior monitoring device for patrol personnel, the training process of the patrol personnel identification model is as follows: Collect continuous historical video frames containing patrol personnel, and mark the position of patrol personnel in each video frame in the historical video frames to obtain a target template, and form a training data set with the target template and the corresponding candidate area selected in the video frame; Design a weight-sharing twin network architecture, the two inputs of which are the template image of the target template and the candidate region of the current video frame, and the output of which is the similarity score; The twin network architecture is iteratively trained using a contrast loss function and a training data set until the network parameters tend to be stable to obtain the patrol personnel recognition model.

[0040] Furthermore, in the above-mentioned patrol behavior monitoring device for patrol personnel, the formula of the contrast loss function is: ; in, N is the number of samples, y i It's a label. y i = 1 means the two input images are similar, y i = 0 means the two input images are not similar, D is the distance between the current candidate region and the target template, m is a hyperparameter representing the minimum distance between non-similar samples. i Indicates i samples. Furthermore, in the above-mentioned patrol behavior monitoring device for patrol personnel, the target patrol path information includes the location information of each point on the patrol path of each patrol area and the corresponding arrival time and stay time, and when the patrol paths included in the target patrol path information meet the preset patrol requirements, the step of determining whether the patrol behavior of the target patrol personnel is in place includes: Using the position information of each point on the patrol path as the ordinate and the corresponding arrival time as the abscissa, the patrol path is mapped to a preset coordinate system to obtain a patrol path line; Obtain the target position and corresponding dwell time of the patrol route, determine the corresponding target point according to the dwell time based on the preset rules, and highlight the target point on the patrol route to obtain the target patrol route; The target patrol route is compared with a standard patrol route, and when the similarity between the target patrol route and the standard patrol route is higher than a threshold, it is determined that the patrol behavior of the target patrol person is in place.

[0041] Furthermore, in the above-mentioned patrol behavior monitoring device for patrol personnel, the step of comparing the target patrol path line with the standard patrol path line and determining that the patrol behavior of the target patrol personnel is in place when the similarity between the target patrol path line and the standard patrol path line is higher than a threshold value includes: Intercepting a preset portion from the target patrol route line and comparing it with a corresponding portion of the standard patrol route line; After the comparison is passed, the remaining target patrol path line is compared with the standard patrol path line, and the patrol behavior of the target patrol person is determined after the remaining target patrol path line is compared with the standard patrol path line.

[0042] Embodiment 4 Another aspect of the present invention further provides a readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the steps of the method described in any one of the above-mentioned embodiments 1 to 2 are implemented.

[0043] Embodiment 5 Another aspect of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the program, the steps of the method described in any one of the above-mentioned embodiments 1 to 2 are implemented.

[0044] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0045] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0046] More specific examples (a non-exhaustive list) of computer-readable storage media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable storage medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0047] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0048] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0049] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for monitoring patrol behavior of patrol personnel, characterized in that: The method is used to monitor the patrol behavior of patrol personnel in the patrol area, and each patrol aisle in the patrol area is equipped with an image acquisition device, and the method includes: When a target patrol officer is identified to enter the patrol area, the patrol route of the target patrol officer is tracked and recorded, and a unique identification code ReID code of the identity characteristics of the target patrol officer in the patrol area is generated through a body feature recognition algorithm; Acquire the patrol information of all patrol areas collected by the image acquisition device within a set time, and analyze the patrol information of each patrol area to obtain the unique identification ReID code of the patrol personnel of each patrol area and the corresponding patrol path information; The target patrol path information of the target patrol person in all patrol areas is obtained by matching the unique identification ReID code of the identity feature of the target patrol person with the unique identification ReID code of the patrol person; When the patrol paths included in the target patrol path information all meet the preset patrol requirements, it is determined that the patrol behavior of the target patrol personnel is in place; The target patrol path information includes the location information of each point on the patrol path of each patrol area and the corresponding arrival time and stay time. When the patrol paths included in the target patrol path information meet the preset patrol requirements, the step of determining whether the patrol behavior of the target patrol personnel is in place includes: Using the position information of each point on the patrol path as the ordinate and the corresponding arrival time as the abscissa, the patrol path is mapped to a preset coordinate system to obtain a patrol path line; Obtain the target position and corresponding dwell time of the patrol route, determine the corresponding target point according to the dwell time based on the preset rules, and highlight the target point on the patrol route to obtain the target patrol route; The target patrol route is compared with a standard patrol route, and when the similarity between the target patrol route and the standard patrol route is higher than a threshold, it is determined that the patrol behavior of the target patrol person is in place.

2. The method for monitoring patrol behavior of patrol personnel according to claim 1, characterized in that: The steps of tracking and recording the patrol route of the target patrol personnel and generating a unique identification ReID code of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm include: When capturing a preset road section in a patrol area, a patrol video of a target patrol person patrolling the preset road section; The patrol video is analyzed to determine the gait and appearance characteristics of the target patrol personnel, and a unique identification ReID code of the identity characteristics of the target patrol personnel is generated according to the gait and appearance characteristics of the target patrol personnel.

3. The method for monitoring patrol behavior of patrol personnel according to claim 1, characterized in that: The patrol information is a patrol video. The steps of acquiring the patrol information of all patrol areas collected by the image acquisition device within a set time, and analyzing the patrol information of each patrol area to obtain a unique identification ReID code of the patrol personnel of each patrol area and corresponding patrol path information include: Decompose the patrol video into continuous video frames, obtain the first video frame where the patrol person appears, identify the first video frame to obtain the initial position of the patrol person, and mark the area of ​​the patrol person to obtain a target template; In each subsequent video frame of the first video frame, the target template and the trained patrol personnel recognition model are used to identify the video frame to obtain subsequent position information of the patrol personnel; The position information of the patrol personnel is continuously recorded, and the time spent at each location is obtained to generate the patrol path information of the patrol personnel.

4. The method for monitoring patrol behavior of patrol personnel according to claim 3, characterized in that: The training process of the patrol personnel identification model is as follows: Collect continuous historical video frames containing patrol personnel, and mark the position of patrol personnel in each video frame in the historical video frames to obtain a target template, and form a training data set with the target template and the corresponding candidate area selected in the video frame; Design a weight-sharing twin network architecture, the two inputs of which are the template image of the target template and the candidate region of the current video frame, and the output of which is the similarity score; The twin network architecture is iteratively trained using a contrast loss function and a training data set until the network parameters tend to be stable to obtain the patrol personnel recognition model.

5. The method for monitoring patrol behavior of patrol personnel according to claim 4, characterized in that: The formula of the contrast loss function is: ； in, N is the number of samples, y i It's a label. y i = 1 means the two input images are similar, y i = 0 means the two input images are not similar, D is the distance between the current candidate region and the target template, m is a hyperparameter representing the minimum distance between non-similar samples. i Indicates i samples.

6. The method for monitoring patrol behavior of patrol personnel according to claim 1, characterized in that: The step of comparing the target patrol route with the standard patrol route and determining that the patrol behavior of the target patrol person is in place when the similarity between the target patrol route and the standard patrol route is higher than a threshold comprises: Intercepting a preset portion from the target patrol route line and comparing it with a corresponding portion of the standard patrol route line; After the comparison is passed, the remaining target patrol path line is compared with the standard patrol path line, and the patrol behavior of the target patrol person is determined after the remaining target patrol path line is compared with the standard patrol path line.

7. A patrol behavior monitoring device for patrol personnel, characterized in that: Used to monitor the patrol behavior of patrol personnel in the patrol area, each patrol aisle in the patrol area is equipped with an image acquisition device, used to implement the patrol behavior monitoring method of patrol personnel according to any one of claims 1 to 6, the device comprising: The recording module is used to track and record the patrol route of the target patrol personnel when it is identified that the target patrol personnel enters the patrol area, and to generate a unique identification code ReID of the identity characteristics of the target patrol personnel in the patrol area through a body feature recognition algorithm; An acquisition module is used to acquire the patrol information of all patrol areas acquired by the image acquisition device within a set time, and analyze the patrol information of each patrol area to obtain a unique identification ReID code of the patrol personnel of each patrol area and corresponding patrol path information; A matching module is used to match the unique identification ReID code of the target patrol personnel's identity characteristics with the unique identification ReID code of the patrol personnel to obtain the target patrol path information of the target patrol personnel in all patrol areas; The monitoring module is used to determine that the patrol behavior of the target patrol personnel is in place when the patrol paths included in the target patrol path information all meet the preset patrol requirements.

8. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 6 when executing the program.

Images