A person monitoring alarm method and system

Abstract

The present application discloses a person monitoring alarm method and system, which uses a camera module to shoot a scene, uses a human target judgment module to judge a human target, uses a human behavior judgment module to determine whether the human target has violated regulations, and uses a display module to mark and display alarms for human targets with violations, thereby achieving real-time monitoring of staff behavior and effectively avoiding staff violations.

Description

Technical Field

[0001] The present invention relates to the technical field of human behavior monitoring, and in particular to a human behavior monitoring alarm method and system. Background Art

[0002] In daily production activities, due to different work contents, the places where staff work often have special requirements. Most places require no smoking and no phone calls. In the prior art, in order to enable staff to comply with the corresponding requirements of the work place, multiple inspectors will be asked to patrol the workplace. Although this way of working can enable staff to do things as required to a certain extent, during the inspection intervals of the inspectors, staff often fail to do things in accordance with the above requirements, which in turn brings certain harm to production activities. In order to avoid similar incidents, there is an urgent need for a method or system that can monitor staff behavior in real time. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for real-time monitoring of staff, so as to realize monitoring of staff's illegal behaviors.

[0004] Therefore, the present invention discloses a person monitoring alarm method, which uses a camera, and the method also includes:

[0005] Acquire the video image captured by the camera, and determine the moving target in the video image based on the multi-moving target detection technology;

[0006] Scan and analyze the moving target according to face recognition technology, and screen and determine the human target;

[0007] Scanning and analyzing the area where the human target is located in the video image to generate a depth image of the human target;

[0008] According to the human skeleton recognition technology, the joint feature data of the human target in the depth image is determined, and the joint feature data is compared with a preset behavior recognition library to determine whether the human target has any violation behavior. If there is any violation behavior, an alarm is issued for the corresponding human target.

[0009] In some embodiments of the present application, in order to be able to judge the behavior of the human target through the joints of the human target, a method for presetting the behavior recognition library is disclosed, and the method for presetting the behavior recognition library includes:

[0010] According to different behaviors of the human body, several human body movements belonging to the human body behaviors are divided, and according to the joint point feature data corresponding to the human body movements, the behavior recognition library associating the human body behaviors with the joint point feature data is generated.

[0011] In some embodiments of the present application, in order to improve the accuracy of the behavior judgment of the human target, the method further includes:

[0012] According to the image edge detection technology, the video image is converted into an edge information graph;

[0013] Determine the face position of the human target according to the face recognition technology, and scan and analyze whether there is a parallel straight line corresponding to the face position of the edge information graph according to the face position;

[0014] If there are parallel straight lines at the position where the face is located in the edge information graph, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

[0015] In some embodiments of the present application, a method for determining whether the edge information graph corresponds to the position of the face with parallel straight lines is disclosed. The method for determining whether the edge information graph corresponds to the position of the face with parallel straight lines includes:

[0016] Performing smoothing and denoising processing on the video image to obtain a denoised image, scanning and analyzing the denoised image, determining high-frequency color features in the denoised image, and thereby generating an edge information map;

[0017] Performing Hough transform on the edge information graph to obtain edge line segment parameters, and determining parallel straight lines and distances between parallel straight lines according to the edge line segment parameters.

[0018] In some embodiments of the present application, in order to determine whether the parallel straight lines at the location of the face are mobile phones or cigarettes, a preset range of distances between the parallel straight lines is provided, and the preset range of distances between the parallel straight lines includes:

[0019] The length and width of a regular mobile phone;

[0020] The diameter of conventional smoke.

[0021] The present application discloses a person monitoring and alarm method, which performs multi-target detection on video images and uses face recognition technology to determine human targets, and compares the joint feature data of the human target with a preset behavior recognition library. If the behavior corresponding to the joint feature data of the human target in the behavior recognition library is a violation, an alarm is issued, thereby achieving real-time monitoring of staff behavior and effectively avoiding staff violations.

[0022] In some embodiments of the present application, a person monitoring alarm system is also disclosed, the system comprising:

[0023] A camera module, used to shoot the scene;

[0024] A display module, used to display the video image captured by the camera;

[0025] A human target determination module determines the moving targets in the video image based on the multi-moving target detection technology, and screens and determines the human targets based on the face recognition technology;

[0026] An image conversion module, used for converting the video image into a depth image;

[0027] A human behavior judgment module, which determines the joint feature data of the human target in the depth image based on the human skeleton recognition technology, and compares the joint feature data with a preset behavior recognition library to determine whether the human target has any violation behavior. If there is any violation behavior, the display module will mark and display the corresponding human target to alarm.

[0028] In some embodiments of the present application, in order to enable the human behavior determination module to determine the behavior of the human target through the behavior recognition library, the content of the behavior recognition library includes:

[0029] Several human behaviors and corresponding joint feature data;

[0030] The correlation between the human behavior and the joint feature data is:

[0031] The human behavior is associated with a number of human movements, and the human movements correspond one-to-one to the corresponding joint feature data.

[0032] In some embodiments of the present application, in order to improve the accuracy of the behavior of the human target of the human behavior judgment module, the method of the human behavior judgment module for identifying the behavior of the human target further includes:

[0033] The image conversion module converts the video image into an edge information graph according to the image edge detection technology;

[0034] The human behavior judgment module determines the face position of the human target according to the face recognition technology, and scans and analyzes whether there are parallel straight lines corresponding to the face position of the edge information graph according to the face position; if there are parallel straight lines corresponding to the face position of the edge information graph, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

[0035] In some embodiments of the present application, in order to enable the system to determine whether there are parallel straight lines at the location of the face, a determination method is disclosed. The method by which the system determines whether there are parallel straight lines at the location of the face corresponding to the edge information graph includes:

[0036] The image conversion module performs smoothing and denoising processing on the video image to obtain a denoised image, performs scanning and analysis on the denoised image, determines high-frequency color features in the denoised image, and generates an edge information map;

[0037] The human behavior judgment module performs Hough transform on the edge information graph to obtain edge line segment parameters, and determines parallel straight lines and distances between parallel straight lines according to the edge line segment parameters.

[0038] In some embodiments of the present application, a content of a preset range of the distance between the parallel straight lines is disclosed, and the preset range of the distance between the parallel straight lines includes:

[0039] The length and width of a regular mobile phone;

[0040] The diameter of conventional smoke.

[0041] The present application discloses a person monitoring and alarm system, which uses a camera module to shoot a scene, uses a human target judgment module to judge a human target, uses a human behavior judgment module to determine whether the human target has violated regulations, and uses a display module to mark and display alarms for human targets with violations, thereby achieving real-time monitoring of staff behavior and effectively avoiding staff violations.

[0042] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a conceptual step for constructing a person monitoring alarm system in the implementation of this application. DETAILED DESCRIPTION

[0044] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0045] Unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0046] Example:

[0047] Therefore, the present invention discloses a person monitoring alarm method, which uses a camera, and the method also includes:

[0048] Step 1: Acquire the video image captured by the camera, and determine the moving target in the video image based on the multi-moving target detection technology.

[0049] Step 2: Scan and analyze the moving target based on face recognition technology to screen and determine the human target.

[0050] Step 3: Scan and analyze the area where the human target is located in the video image to generate a depth image of the human target.

[0051] Step 4: Determine the joint feature data of the human target in the depth image based on the human skeleton recognition technology.

[0052] Step 5, comparing the joint feature data with a preset behavior recognition library to determine whether the human target has any illegal behavior. If there is any illegal behavior, an alarm is issued for the corresponding human target.

[0053] In step 5, in order to be able to judge the behavior of the human target through the joints of the human target, in some embodiments of the present application, a method for presetting the behavior recognition library is disclosed, and the method for presetting the behavior recognition library includes: according to different human behaviors, dividing a number of human body movements belonging to human behaviors, and according to the joint feature data corresponding to the human body movements, generating the behavior recognition library that associates the human behavior with the joint feature data.

[0054] In some embodiments of the present application, in order to improve the accuracy of the behavior judgment of the human target, the method further includes:

[0055] Step 6, converting the video image into an edge information graph according to image edge detection technology;

[0056] Step 7, determining the face position of the human target according to the face recognition technology, and scanning and analyzing whether there are parallel straight lines corresponding to the face position of the edge information graph according to the face position;

[0057] Step 8: If there are parallel straight lines at the position where the face is located in the edge information graph, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

[0058] In some embodiments of the present application, a method for determining whether the edge information graph corresponds to the position of the face with parallel straight lines is disclosed. The method for determining whether the edge information graph corresponds to the position of the face with parallel straight lines includes:

[0059] a. performing smoothing and denoising processing on the video image to obtain a denoised image, scanning and analyzing the denoised image, determining high-frequency color features in the denoised image, and thereby generating an edge information map;

[0060] b. Performing Hough transform on the edge information graph to obtain edge line segment parameters, and determining parallel straight lines and distances between parallel straight lines based on the edge line segment parameters.

[0061] In some embodiments of the present application, in order to determine whether the parallel straight lines at the location of the face are mobile phones or cigarettes, a preset range of distances between the parallel straight lines is provided, and the preset range of distances between the parallel straight lines includes:

[0062] 1. The length and width of a regular mobile phone.

[0063] 2. The diameter of conventional smoke.

[0064] The present application discloses a person monitoring and alarm method, which performs multi-target detection on video images and uses face recognition technology to determine human targets, and compares the joint feature data of the human target with a preset behavior recognition library. If the behavior corresponding to the joint feature data of the human target in the behavior recognition library is a violation, an alarm is issued, thereby achieving real-time monitoring of staff behavior and effectively avoiding staff violations.

[0065] In order to further explain the technical solution of the present application, the present application also discloses a conceptual step for constructing a person monitoring alarm system. First of all, it should be made clear that the technical problem that the person monitoring alarm system needs to solve is to monitor whether the staff has any violations, and the violations include smoking and using mobile phones.

[0066] The ideation steps are as follows:

[0067] S1, obtaining the video image captured by the camera, and determining the moving target in the video image based on the multi-moving target detection technology.

[0068] S2, scanning and analyzing the moving target based on face recognition technology, and screening and determining the human target.

[0069] S3, scanning and analyzing the area where the human target is located in the video image to generate a depth image of the human target.

[0070] S4, determining the joint feature data of the human target in the depth image according to the human skeleton recognition technology.

[0071] S5, comparing the joint feature data with a preset behavior recognition library to determine whether the human target has any illegal behavior.

[0072] S6: If there is any violation, an alarm is issued for the corresponding human target.

[0073] Among them, in the step S5, the behaviors recorded by the behavior recognition library are divided into smoking behaviors and mobile phone use behaviors. According to the action characteristics of these two behaviors, the human body target joint feature data is obtained, and the joint feature data includes the position relationship between the joints and the time to maintain the position relationship. Although the smoking behavior and mobile phone use behaviors disclosed here, there are still many illegal behaviors that have their own action characteristics. If you want to monitor other illegal behaviors, you can still use the above means to set the behavior recognition library.

[0074] The method for presetting the behavior recognition library includes dividing a number of human body actions belonging to human behaviors according to different human body behaviors, and generating the behavior recognition library that associates human behaviors with the joint point feature data according to the joint point feature data corresponding to the human body actions.

[0075] In order to improve the accuracy of the system monitoring behavior, the face of the human target can also be scanned and monitored to determine whether there is a mobile phone and a cigarette. The cross-sections of these two objects are both rectangular. Compared with the human face, these two objects have parallel edges. Based on this feature, some embodiments of the present application also disclose the following conceptual steps:

[0076] S7, converting the video image into an edge information graph according to image edge detection technology.

[0077] S8, determining the face position of the human target according to the face recognition technology.

[0078] S8, based on the location of the face, scan and analyze whether there are parallel straight lines corresponding to the location of the face in the edge information graph.

[0079] S9, if there are parallel straight lines at the position where the edge information graph corresponds to the face, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

[0080] In some embodiments of the present application, a person monitoring alarm system is also disclosed, which includes a camera module, a display module, a human target judgment module, an image conversion module and a human behavior judgment module.

[0081] The camera module is used for shooting the scene.

[0082] The display module is used to display the video image captured by the camera.

[0083] The human target judgment module determines the moving targets in the video image based on the multi-moving target detection technology, and screens and determines the human targets based on the face recognition technology.

[0084] The image conversion module is used to convert the video image into a depth image.

[0085] The human behavior judgment module determines the joint point feature data of the human target in the depth image based on the human skeleton recognition technology, and compares the joint point feature data with a preset behavior recognition library to determine whether the human target has any violation behavior. If there is any violation behavior, the display module marks and displays an alarm for the corresponding human target.

[0086] In some embodiments of the present application, in order to enable the human behavior determination module to determine the behavior of the human target through the behavior recognition library, the content of the behavior recognition library includes:

[0087] Several human behaviors and their corresponding joint feature data.

[0088] The correlation between the human behavior and the joint feature data is:

[0089] The human behavior is associated with a number of human movements, and the human movements correspond one-to-one to the corresponding joint feature data.

[0090] In some embodiments of the present application, in order to improve the accuracy of the behavior of the human target of the human behavior judgment module, the method of the human behavior judgment module for identifying the behavior of the human target further includes:

[0091] a. The image conversion module converts the video image into an edge information map according to the image edge detection technology;

[0092] b. The human behavior judgment module determines the face position of the human target according to the face recognition technology, and scans and analyzes whether there are parallel straight lines corresponding to the face position of the edge information graph according to the face position; if there are parallel straight lines corresponding to the face position of the edge information graph, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

[0093] In some embodiments of the present application, in order to enable the system to determine whether there are parallel straight lines at the location of the face, a determination method is disclosed. The method by which the system determines whether there are parallel straight lines at the location of the face corresponding to the edge information graph includes:

[0094] a. The image conversion module performs smoothing and denoising processing on the video image to obtain a denoised image, scans and analyzes the denoised image, determines the high-frequency color features in the denoised image, and generates an edge information map;

[0095] b. The human behavior judgment module performs Hough transform on the edge information graph to obtain edge line segment parameters, and determines parallel straight lines and distances between parallel straight lines based on the edge line segment parameters.

[0096] In some embodiments of the present application, a content of a preset range of the distance between the parallel straight lines is disclosed, and the preset range of the distance between the parallel straight lines includes:

[0097] The length and width of a regular mobile phone;

[0098] The diameter of conventional smoke.

[0099] The present application discloses a person monitoring and alarm system, which uses a camera module to shoot a scene, uses a human target judgment module to judge a human target, uses a human behavior judgment module to determine whether the human target has violated regulations, and uses a display module to mark and display alarms for human targets with violations, thereby achieving real-time monitoring of staff behavior and effectively avoiding staff violations.

[0100] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A person monitoring alarm method, characterized in that: The method comprises: Acquire a video image captured by a camera, and determine a moving target in the video image based on a multi-moving target detection technology; Scan and analyze the moving target according to face recognition technology to screen and determine the human target; Scanning and analyzing the area where the human target is located in the video image to generate a depth image of the human target; Determine the joint feature data of the human target in the depth image according to the human skeleton recognition technology, and compare the joint feature data with a preset behavior recognition library to determine whether the human target has any illegal behavior. If there is any illegal behavior, an alarm is issued for the corresponding human target; Also includes: According to the image edge detection technology, the video image is converted into an edge information graph; Determine the face position of the human target according to the face recognition technology, and scan and analyze whether there is a parallel straight line corresponding to the face position of the edge information graph according to the face position; If there are parallel straight lines at the position of the face corresponding to the edge information graph, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

2. A person monitoring alarm method according to claim 1, characterized in that: The method of presetting the behavior recognition library includes: According to different behaviors of the human body, several human body movements belonging to the human body behaviors are divided, and according to the joint point feature data corresponding to the human body movements, the behavior recognition library associating the human body behaviors with the joint point feature data is generated.

3. A person monitoring alarm method according to claim 1, characterized in that: The method for determining whether the edge information graph corresponds to the position where the face is located includes: Performing smoothing and denoising processing on the video image to obtain a denoised image, scanning and analyzing the denoised image, determining high-frequency color features in the denoised image, and thereby generating an edge information map; Performing Hough transform on the edge information graph to obtain edge line segment parameters, and determining parallel straight lines and distances between parallel straight lines according to the edge line segment parameters.

4. A person monitoring alarm method according to claim 1, characterized in that: The preset range of the distance between the parallel straight lines includes: The length and width of a regular mobile phone; The diameter of conventional smoke.

5. A person monitoring alarm system, characterized in that: The system comprises: A camera module, used to shoot the scene; A display module, used to display the video image captured by the camera module; A human target determination module determines the moving targets in the video image based on the multi-moving target detection technology, and screens and determines the human targets based on the face recognition technology; An image conversion module, used for converting the video image into a depth image; A human behavior judgment module, which determines the joint feature data of the human target in the depth image based on the human skeleton recognition technology, and compares the joint feature data with a preset behavior recognition library to determine whether the human target has any illegal behavior. If there is any illegal behavior, the display module marks and displays an alarm on the corresponding human target; The method for the human behavior judgment module to identify the human target behavior also includes: The image conversion module converts the video image into an edge information graph according to the image edge detection technology; The human behavior judgment module determines the face position of the human target according to the face recognition technology, and scans and analyzes whether there are parallel straight lines corresponding to the face position of the edge information graph according to the face position; if there are parallel straight lines corresponding to the face position of the edge information graph, and the distance between the parallel straight lines falls within a preset range, it is determined that the corresponding human target has violated the rules, and an alarm is issued for the corresponding human target.

6. A person monitoring alarm system according to claim 5, characterized in that: The content of the behavior recognition library includes: Several human behaviors and corresponding joint feature data; The correlation between the human behavior and the joint feature data is: The human behavior is associated with a number of human movements, and the human movements correspond one-to-one to the corresponding joint feature data.

7. A person monitoring alarm system according to claim 5, characterized in that: The method for the system to determine whether the edge information graph corresponds to the position where the face is located includes: The image conversion module performs smoothing and denoising processing on the video image to obtain a denoised image, performs scanning and analysis on the denoised image, determines high-frequency color features in the denoised image, and generates an edge information map; The human behavior judgment module performs Hough transform on the edge information graph to obtain edge line segment parameters, and determines parallel straight lines and distances between parallel straight lines according to the edge line segment parameters.

8. The person monitoring alarm system according to claim 5, characterized in that: The preset range of the distance between the parallel straight lines includes: The length and width of a regular mobile phone; The diameter of conventional smoke.

Images