A method for identifying the behavior of people kicking on a mine car

Abstract

The present invention relates to the field of behavior recognition technology, specifically to a method for identifying the behavior of people kicking a mine car, the method comprising: demarcating the mine car operation area to obtain a target area, marking the people in the target area as target people; collecting the behavior action data and action trajectory data of the target people, obtaining the overlap degree of abnormal behavior characteristics and the relative position risk assessment index of each target person through processing, comprehensively analyzing to obtain the abnormal behavior risk assessment index of each target person, conducting risk assessment on the abnormal behavior of the people according to the abnormal behavior risk assessment index of each target person, and feeding back the assessment results. The present invention conducts a comprehensive analysis from the two dimensions of behavior and position, can more comprehensively capture the abnormal behavior characteristics of the target people, make the risk assessment more comprehensive and accurate, and divide the target people into risk levels, so as to facilitate the adoption of corresponding management measures, which is conducive to improving the pertinence and efficiency of management.

Description

Technical Field

[0001] The present invention relates to the technical field of behavior recognition, and in particular to a method for recognizing the behavior of a person leaning on a mine car. Background Art

[0002] In coal mines and other mining production environments, people stepping on the mine car is an extremely dangerous and unsafe behavior, which may not only endanger personal life safety, but also cause serious production accidents. Therefore, it is necessary to identify the behavior of people stepping on the mine car to achieve real-time monitoring and early warning of miners' unsafe behavior.

[0003] For example, the invention patent with the announcement number: CN102227616B is a behavior recognition device, which includes: a hand movement observation unit, which outputs hand movement data representing the movement of a user's hand; a speed change detection unit, which detects the moment when the user's hand movement adopts a prescribed behavior pattern based on the hand movement data; a feature selection unit, which selects the hand movement data output at the moment detected by the speed change detection unit; a behavior standard pattern storage unit, which stores a sequence of action pattern data represented by a sequence of hand movement data according to each behavior of the user; and a behavior recognition unit, which recognizes the behavior represented by the action pattern data sequence with the greatest similarity as the behavior of the user, and the similarity is calculated according to each action pattern data sequence stored in the behavior standard pattern storage unit, based on the action pattern data sequence and the hand movement data selected by the feature selection unit.

[0004] For example, the invention patent with announcement number: CN113405580B is a smoking behavior recorder with automatic ash ejection, including a recorder body, a power module and a control module arranged in the recorder body; a filter rod insertion channel is provided on the recorder body; a vibration module is provided on the outer wall of the filter rod insertion channel, the power module is connected to the vibration module through a power supply line, a switch device is provided on the power supply line, and the control module controls the control end connected to the switch device to realize the on-off control of the switch device; it also includes an image acquisition module and an image recognition module connected to the image acquisition module, the image acquisition area of ​​the image acquisition module overlaps with the front area of ​​the orifice of the filter rod insertion channel; the image recognition module is connected to the control module. When it is necessary to eject ash, there is no need to touch the cigarette or shake the recorder, the recorder can automatically vibrate the cigarette through the vibration device when needed to make the ash fall through image recognition.

[0005] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, it was found that the above technology has at least the following technical problems: the current behavior recognition method generally uses image recognition, which often identifies and analyzes behaviors that have already occurred. There may be a delay between data collection and behavior recognition, and the occurrence of dangerous behaviors cannot be stopped in time, posing a safety hazard. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a method for identifying the behavior of people leaning on a mine car, which can effectively solve the problems involved in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The present invention provides a method for identifying the behavior of people kicking a mine car, comprising: demarcating the mine car operation area to obtain a target area, and marking the people in the target area as target people.

[0008] The behavioral action data of the target personnel are collected and processed to obtain the overlap of abnormal behavioral characteristics of each target person.

[0009] The target personnel's movement trajectory data is collected and processed to obtain the relative position risk assessment index of each target personnel.

[0010] Based on the overlap of abnormal behavior characteristics of each target person and the relative position risk assessment index, a comprehensive analysis is conducted to obtain the abnormal behavior risk assessment index of each target person. Based on the abnormal behavior risk assessment index of each target person, a risk assessment of the abnormal behavior of the personnel is conducted and the assessment results are fed back.

[0011] As a further method, the target area is obtained by delineating the mine car operation area. The specific process is: extracting the minimum safe distance between personnel and the mine car from the mine car database, taking the mine car position point as the center and the minimum safe distance between personnel and the mine car as the radius to delineate the area, marking the delineated area as the target area, and marking the personnel in the target area as the target personnel.

[0012] As a further method, the target person's behavior and action data are collected, and the specific process is: collecting real-time video image data of the target person within a preset time period, including the target person's movement trajectory data and behavior and action data within the preset time period, and the target person's behavior and action data includes the target person's behavior feature image at each instantaneous recording point.

[0013] As a further method, the above-mentioned processing obtains the overlap degree of abnormal behavior characteristics of each target person. The specific processing process is: according to the behavior characteristic image of each target person at each instantaneous recording point, a number of key joint position points are deployed on the image, and a two-dimensional coordinate system is constructed to mark each key joint position point, which is recorded as: , extract each reference abnormal behavior feature image from the mine car database, match the similar abnormal behavior feature image according to the behavior feature image of each target person at each instantaneous recording point, and collect each key joint position point of the similar abnormal behavior feature image, recorded as , and the overlap of abnormal behavior characteristics of each target person is obtained through comprehensive analysis.

[0014] As a further method, the movement trajectory data of the target personnel include the distance between each target personnel and the mine car and the movement speed of each target personnel within a preset time period.

[0015] As a further method, the relative position risk assessment index of each target person is obtained through processing. The specific process is: marking the distance between each target person and the mine car as the relative distance of each target person, extracting the relative distance of each target person between the previous adjacent instantaneous recording point and the current time point from the movement trajectory data of each target person within a preset time period, and subtracting the relative distance of each target person between the previous adjacent instantaneous recording point and the current time point. If the difference is greater than 0, it indicates that the target person is gradually approaching the mine car. If the difference is less than 0, it indicates that the target person is gradually moving away from the mine car.

[0016] According to the movement speed and relative distance of each target person within a preset time period, the critical movement speed and critical relative distance are extracted from the mine car database, and the relative position risk assessment index of each target person is obtained through comprehensive analysis.

[0017] As a further method, the comprehensive analysis obtains the abnormal behavior risk assessment index of each target person. The specific analysis process is: based on the overlap of abnormal behavior characteristics of each target person and the relative position risk assessment index, a comprehensive analysis is performed to obtain the abnormal behavior risk assessment index of each target person. The abnormal behavior risk assessment index of each target person is used to quantitatively assess the risk level of abnormal behavior of each target person.

[0018] As a further method, the risk assessment of the behavior of personnel leaning on the mine car is performed according to the abnormal behavior risk assessment index of each target person. The specific assessment process is: according to the abnormal behavior risk assessment index of each target person, the abnormal behavior risk assessment index threshold of the target person is extracted from the mine car database, and the abnormal behavior risk assessment index of each target person is compared with the abnormal behavior risk assessment index threshold of the target person. If the abnormal behavior risk assessment index of a target person is higher than the abnormal behavior risk assessment index threshold of the target person, it indicates that the risk of abnormal behavior of the target person is relatively large, and the target person is marked as a high-risk person. If the abnormal behavior risk assessment index of a target person is equal to the abnormal behavior risk assessment index threshold of the target person, it indicates that the risk of abnormal behavior of the target person is at a medium level, and the target person is marked as a medium-risk person. If the abnormal behavior risk assessment index of a target person is lower than the abnormal behavior risk assessment index threshold of the target person, it indicates that the risk of abnormal behavior of the target person is relatively small, and the target person is marked as a low-risk person, and the assessment result is fed back.

[0019] As a further method, the relative position risk assessment index of each target person is specifically expressed as:

[0020] ;

[0021] In the formula, It represents the relative position risk assessment index of the nth target person, and n represents the number of each target person. , m represents the total number of target personnel, Indicates the action speed of the nth target person, Indicates the relative distance of the nth target person, represents the critical action speed, represents the critical relative distance, Indicates the relative distance of the nth target person at the previous adjacent instantaneous recording point, Indicates the relative distance of the nth target person at the current time point, Indicates the relative position risk assessment impact factor corresponding to the preset approach action speed, Indicates the relative position risk assessment impact factor corresponding to the preset relative distance when approaching. Indicates the relative position risk assessment impact factor corresponding to the preset moving speed when moving away. Indicates the relative position risk assessment impact factor corresponding to the preset relative distance.

[0022] As a further method, the specific numerical expression of the abnormal behavior risk assessment index of each target person is:

[0023] ;

[0024] In the formula, represents the abnormal behavior risk assessment index of the nth target person, Indicates the overlap degree of abnormal behavior characteristics of the nth target person, Represents the relative position risk assessment index of the nth target person, Indicates the abnormal behavior risk assessment impact factor corresponding to the preset abnormal behavior feature overlap, Indicates the abnormal behavior risk assessment impact factor corresponding to the preset relative position risk assessment index.

[0025] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0026] (1) The present invention provides a method for identifying the behavior of people leaning on a mine car, analyzes and obtains the abnormal behavior risk assessment index of each target person, and conducts a comprehensive analysis from the two dimensions of behavior and position. It can more comprehensively capture the abnormal behavior characteristics of the target person, making the risk assessment more comprehensive and accurate, and classifying the target persons into risk levels, so that management personnel can take corresponding management measures according to the risk level, which is conducive to improving the pertinence and efficiency of management. At the same time, it can also achieve reasonable allocation of resources to ensure that high-risk areas and high-risk personnel receive more attention and resource investment. This is conducive to reducing the risk of abnormal behavior of target personnel and improving the safety and efficiency of mine operations;

[0027] (2) By defining the target area, the present invention can ensure that no person is allowed to approach the dangerous distance without authorization within the mine car operation area. This is conducive to reducing the risk of accidental injury caused by people being too close to the mine car. Once the system identifies the target person, it can immediately trigger the early warning mechanism to remind the relevant personnel to pay attention to safety. This instant feedback mechanism helps to quickly respond to and prevent potential dangerous behaviors;

[0028] (3) The present invention automatically identifies similar or identical abnormal behavior patterns by matching the target person's behavior feature image with the reference abnormal behavior feature image in the mine car database. This helps to improve the accuracy and efficiency of recognition and reduce the subjectivity and uncertainty of human judgment. Comprehensive analysis of the overlap of the abnormal behavior features of each target person can further confirm whether there is a risk of pedaling the mine car. The higher the overlap, the more similar the target person's behavior is to the known behavior features of pedaling the mine car, which helps to improve the reliability and accuracy of recognition;

[0029] (4) The present invention can realize dynamic tracking of personnel behavior by continuously monitoring the action trajectory data of each target person within a preset time period and analyzing the relative position risk assessment index of each target person. This real-time performance helps to timely discover and warn of possible risk behaviors, and can also accurately assess the risk status of each target person relative to the mine car, thereby identifying which personnel may be engaging in or preparing to engage in dangerous behaviors of kicking the mine car. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The present invention is further described with the aid of the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other accompanying drawings based on the following accompanying drawings without creative work.

[0031] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0032] Figure 2It is a schematic diagram of the functional relationship between the overlap degree of abnormal behavior characteristics of the target person and the abnormal behavior risk assessment index involved in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0034] Reference Figure 1 As shown, the present invention provides a method for identifying the behavior of people kicking a mine car, comprising: demarcating the mine car operation area to obtain a target area, and marking people in the target area as target people.

[0035] Specifically, the mine car operation area is delineated to obtain the target area. The specific process is: extract the minimum safe distance between personnel and the mine car from the mine car database, take the mine car position point as the center, and use the minimum safe distance between personnel and the mine car as the radius to delineate the area, mark the delineated area as the target area, and mark the personnel in the target area as target personnel.

[0036] In a specific embodiment, by setting the minimum safe distance between personnel and mine cars as the standard for demarcating areas, it can be ensured that no personnel in the mine car operation area can approach this dangerous distance without authorization. This directly reduces the risk of accidental injury caused by personnel being too close to the mine car, especially the highly dangerous behavior of leaning on the mine car. Once the system identifies the presence of personnel, namely the target personnel, in the target area, it can immediately trigger the early warning mechanism to remind the target personnel to pay attention to safety. This instant feedback mechanism helps to quickly respond to and prevent potential dangerous behaviors, such as leaning on the mine car. By clearly demarcating the mine car operation area, personnel and mine cars can be guided to operate according to established paths and rules, reducing confusion and conflicts. This not only improves work efficiency, but also creates a more orderly and safe working environment. By continuously collecting and analyzing data, safety strategies can be further optimized and the accuracy and effectiveness of risk identification can be improved. The long-term implementation of this safety management measure based on demarcated areas can enable employees to gradually develop the habit of complying with safety regulations, improve overall safety awareness, and actively avoid dangerous behaviors such as leaning on the mine car. In an emergency, such as when the mine car is out of control or fails, the system can quickly determine the distribution of personnel in the target area and provide important information for emergency response. This helps to evacuate people quickly and reduce accident losses.

[0037] The behavioral action data of the target personnel are collected and processed to obtain the overlap of abnormal behavioral characteristics of each target person.

[0038] Specifically, the target person's behavior and action data are collected, and the specific process is: collecting real-time video image data of the target person within a preset time period, including the target person's motion trajectory data and behavior and action data within the preset time period, and the target person's behavior and action data includes the target person's behavior feature image at each instantaneous recording point.

[0039] It should be explained that the behavior and action data of the target person also includes the thermal imaging image of the target person.

[0040] Furthermore, the overlap of abnormal behavior characteristics of each target person is obtained through processing. The specific processing process is as follows: according to the behavior characteristic image of each target person at each instantaneous recording point, several key joint position points are deployed on the image, and a two-dimensional coordinate system is constructed to mark each key joint position point, which is recorded as:

[0041] , extract each reference abnormal behavior feature image from the mine car database, match the similar abnormal behavior feature image according to the behavior feature image of each target person at each instantaneous recording point, and collect each key joint position point of the similar abnormal behavior feature image, which is recorded as: ,The overlap of abnormal behavior characteristics of each target person is obtained through comprehensive analysis;

[0042] It needs to be explained that key joint locations refer to the locations of joints that play a key role in connection and movement in the human body, specifically including shoulder joints, elbow joints, ankle joints, knee joints, hip joints and wrist joints.

[0043] Specifically, according to the behavior feature images of each target person at each instantaneous recording point, the process of matching to obtain similar abnormal behavior feature images is as follows: according to the key joint position points of each target person, the joint connection line of the target person is constructed, and then the joint connection line corresponding to the reference joint connection line of various abnormal behavior categories stored in the mine car database is overlapped and compared, and the length of the overlapped joint connection line is extracted. , and extract the reference joint connection line lengths corresponding to various abnormal behavior categories , calculate the matching degree between the target person and each behavior category , and its calculation formula is: , extract the behavior feature image of the target person in the abnormal behavior category with the largest matching degree, and mark it as a similar abnormal behavior feature image;

[0044] It should be understood that the overlap of abnormal behavior characteristics of each target person is a quantitative evaluation data obtained through comprehensive analysis of the key joint position points of each target person, which is used to quantitatively evaluate the similarity between the behavior characteristics of each target person and similar abnormal behavior characteristics, and provide a basis for the assessment of personnel behavior risks.

[0045] In a specific embodiment, the numerical expression of the overlap of abnormal behavior characteristics of each target person is:

[0046] ;

[0047] In the formula, represents the overlap of abnormal behavior characteristics of the nth target person, n represents the number of each target person, , m represents the total number of target personnel, i represents the number of each key joint, , s represents the total number of key joints, r represents the number of each instantaneous recording point, , h represents the total number of instantaneous recording points, Represents the horizontal coordinate of the i-th key joint position point of the n-th target person at the r-th instantaneous recording point, It represents the ordinate of the i-th key joint position point of the n-th target person at the r-th instantaneous recording point, Represents the horizontal coordinate of the i-th key joint position point of the reference standard, Represents the ordinate of the i-th key joint position point of the reference standard.

[0048] It needs to be explained that, from the above embodiments, it can be obtained that the smaller the difference between the horizontal and vertical coordinates of the key joint position points of the target person and the horizontal and vertical coordinates of the key joint position points of the reference standard, the greater the overlap of the abnormal behavior characteristics of the target person, which means that the closer the abnormal behavior of the target person is to similar abnormal behavior, the greater the risk of abnormal behavior of the target person.

[0049] It should be explained that in this embodiment, by extracting and marking the key joint position points of each target person and accurately representing them in a two-dimensional coordinate system, subtle behavioral changes can be captured. This high-precision method enables the system to more accurately identify abnormal behavioral features related to lying on the mine car, such as body tilt, abnormal leg position, etc. Based on the analysis of the behavioral feature image of the instantaneous recording point, this process can monitor the behavior of the target person in real time. This helps to timely discover and prevent dangerous behaviors such as lying on the mine car and reduce potential safety risks. By matching the behavioral feature image of the target person with the reference abnormal behavior feature image in the mine car database, the system can automatically identify similar or identical abnormal behavior patterns. This intelligent matching and comparison technology improves the accuracy and efficiency of recognition and reduces the subjectivity and uncertainty of human judgment. Comprehensively analyzing the overlap of the abnormal behavior features of each target person can further confirm whether there is a risk of lying on the mine car. The higher the overlap, the more similar the behavior of the target person is to the known behavior features of the lying on the mine car, thereby improving the reliability and accuracy of recognition. The whole process is based on data analysis and machine learning technology, which can automatically extract useful information from a large amount of data and provide strong support for safety management decisions. Based on the analysis results provided by the system, you can respond quickly and take corresponding safety measures.

[0050] It needs to be explained that the real-time video image data of the target person within a preset time period is often affected by environmental factors, resulting in inaccurate image feedback. The environmental data and image quality feedback data that affect the image quality at each instantaneous recording point are collected, and a comprehensive analysis is performed to obtain the image quality assessment value of the target area.

[0051] Specifically, the process of comprehensive analysis to obtain the image quality assessment value of the target area is as follows: deploy a number of environmental monitoring points, extract the environmental data at the current time point from the environmental data that affects the image quality at each instantaneous recording point, including lighting intensity and dust concentration, extract the reference standard lighting intensity, allowable deviation lighting intensity and critical dust concentration as well as the reference standard pixel value and allowable deviation pixel value from the mine car database, and extract the pixel value of each pixel point in the image at the current time point from the image quality feedback data of each instantaneous recording point, and conduct comprehensive analysis to obtain the image quality assessment value of the target area.

[0052] It should be understood that the target area image quality assessment value is a quantitative assessment data obtained by comprehensively analyzing the illumination intensity, dust concentration and image pixel value at the current time point and the image pixel value at the previous instantaneous recording point, and is used to quantitatively assess the image quality.

[0053] In a specific embodiment, the numerical expression of the target area image quality evaluation value is:

[0054] ;

[0055] In the formula, represents the image quality assessment value of the target area, e represents the natural constant, j represents the number of each environmental monitoring point, , f represents the total number of environmental monitoring points, represents the lighting intensity of the jth environmental monitoring point, represents the dust concentration at the jth environmental monitoring point, Indicates the reference standard lighting intensity, Indicates the allowable deviation lighting intensity, represents the critical dust concentration, k represents the number of each pixel, , q represents the total number of pixels, Represents the pixel value of the kth pixel in the image at the current time point, Indicates the reference standard pixel value, Indicates the allowable deviation pixel value, Indicates the image quality assessment influencing factor corresponding to the preset lighting intensity, Indicates the image quality assessment impact factor corresponding to the preset dust concentration, Indicates the image quality assessment influencing factor corresponding to the preset image pixel value.

[0056] It should be understood that, from the above embodiments, the smaller the difference between the lighting intensity at the current time point and the lighting intensity of the reference standard, the smaller the dust concentration, the smaller the difference between the image pixel value at the current time point and the reference standard image pixel value, the larger the image quality evaluation value, indicating that the image quality is better.

[0057] It should be explained that in this embodiment, the image quality assessment influence factor corresponding to the lighting intensity, the image quality assessment influence factor corresponding to the dust concentration, and the image quality assessment influence factor corresponding to the image pixel value are in the range of 0 to 1. Through the relationship between the lighting intensity, dust concentration, and image pixel value in the historical data and the image quality assessment value of the target area, a mapping set of lighting intensity, dust concentration, and image pixel value and their corresponding image quality assessment influence factors is established. The real-time lighting intensity, dust concentration, and image pixel value are input to obtain the image quality assessment influence factor corresponding to the lighting intensity, the image quality assessment influence factor corresponding to the dust concentration, and the image quality assessment influence factor corresponding to the image pixel value from the mapping set. The above-mentioned influence factors are all extracted from the mine car database.

[0058] It should be explained that the pixel value refers to the brightness of each pixel in the image. Through analysis, we can understand the overall brightness distribution, contrast, and whether there are problems such as overexposure or underexposure of the image. Appropriate lighting intensity helps to ensure the correct exposure of the image and avoid too bright or too dark. Different light sources will affect the color performance of the image. For example, the colors under natural light are usually more natural than those under fluorescent lights. Good lighting can reduce unnecessary shadows and make the image clearer. Dust particles in the air scatter light and reduce the clarity of the image. When the dust concentration is high, the contrast of the image may decrease. Dust can also cause image color shift, especially when it is more obvious during long exposure or using light of a specific wavelength. By analyzing the ambient lighting intensity, the exposure time and gain of the image can be adjusted to obtain better brightness and color balance. Changes in dust concentration will affect the clarity of the image. By analyzing the dust concentration, measures can be taken to reduce or eliminate the impact of dust on image quality. Statistical analysis of pixel values ​​can help identify and reduce noise in images.

[0059] Furthermore, the image is evaluated according to the target area image quality assessment value. The specific process is: extracting the target area image quality assessment threshold from the mine car database, comparing the target area image quality assessment value with the target area image quality assessment threshold, if the target area image quality assessment value is higher than or equal to the target area image quality assessment threshold, then it indicates that the target area image quality is qualified; if the target area image quality assessment value is lower than the target area image quality assessment threshold, then it indicates that the target area image quality is unqualified.

[0060] It should be understood that the real-time video image data of the target person in the above embodiment is obtained by using the high-definition camera deployed in the mining area to collect the image analysis of the target area in real time. If the quality of the collected image of the target area is unqualified, it is necessary to use an infrared thermal imager to collect the thermal imaging image of the target person.

[0061] The target personnel's movement trajectory data is collected and processed to obtain the relative position risk assessment index of each target personnel.

[0062] Specifically, the target personnel's movement trajectory data includes the distance between each target personnel and the mine car and the movement speed of each target personnel within a preset time period.

[0063] Furthermore, the relative position risk assessment index of each target person is obtained after processing. The specific process is: marking the distance between each target person and the mine car as the relative distance of each target person, extracting the relative distance of each target person between the previous adjacent instantaneous recording point and the current time point from the movement trajectory data of each target person within a preset time period, and subtracting the relative distance of each target person between the previous adjacent instantaneous recording point and the current time point. If the difference is greater than 0, it indicates that the target person is gradually approaching the mine car. If the difference is less than 0, it indicates that the target person is gradually moving away from the mine car.

[0064] According to the movement speed and relative distance of each target person within a preset time period, the critical movement speed and critical relative distance are extracted from the mine car database, and the relative position risk assessment index of each target person is obtained through comprehensive analysis.

[0065] It should be understood that the relative position risk assessment index of each target person is a quantitative assessment data obtained through a comprehensive analysis of the movement speed and relative distance of each target person within a preset time period. It is used to quantitatively assess the potential risk level of the relative position of each target person and provide a basis for the assessment of personnel behavior risks.

[0066] In a specific embodiment, the numerical expression of the relative position risk assessment index of each target person is:

[0067] ;

[0068] In the formula, Represents the relative position risk assessment index of the nth target person, Indicates the action speed of the nth target person, Indicates the relative distance of the nth target person, Indicates the critical action speed, represents the critical relative distance, Indicates the relative distance of the nth target person at the previous adjacent instantaneous recording point, Indicates the relative distance of the nth target person at the current time point, Indicates the relative position risk assessment impact factor corresponding to the preset approach action speed, Indicates the relative position risk assessment impact factor corresponding to the preset relative distance when approaching. Indicates the relative position risk assessment impact factor corresponding to the preset moving speed when moving away. Indicates the relative position risk assessment impact factor corresponding to the preset relative distance.

[0069] It should be explained that, in this embodiment, the relative position risk assessment influence factor corresponding to the movement speed when approaching is set to 0.5, the relative position risk assessment influence factor corresponding to the relative distance when approaching is set to 1, the relative position risk assessment influence factor corresponding to the movement speed when moving away is set to 0.8, the relative position risk assessment influence factor corresponding to the relative distance when moving away is set to 0.5, the critical movement speed is set to 1 m / s, and the critical relative distance is set to 10 meters.

[0070] It should be explained that in the above embodiments: , which can be achieved by adjusting the value of the influencing factor.

[0071] It should be explained that, in this embodiment, the relative position risk assessment impact factor corresponding to the approaching action speed and the relative position risk assessment impact factor corresponding to the approaching relative distance have a value range of 0 to 1. Through the relationship between the approaching action speed and the approaching relative distance and the relative position risk assessment index in the historical data, a mapping set of the approaching action speed and the approaching relative distance and their corresponding relative position risk assessment impact factors is established. The real-time approaching action speed and the approaching relative distance are input to obtain the relative position risk assessment impact factor corresponding to the approaching action speed and the relative position risk assessment impact factor corresponding to the approaching relative distance from the mapping set. The above impact factors are all extracted from the mine car database.

[0072] It should be explained that, in this embodiment, the relative position risk assessment impact factor corresponding to the moving speed when moving away and the relative position risk assessment impact factor corresponding to the relative distance when moving away are in the range of 0 to 1. Through the relationship between the moving speed when moving away and the relative distance when moving away and the relative position risk assessment index in the historical data, a mapping set of the moving speed when moving away and the relative distance when moving away and their corresponding relative position risk assessment impact factors is established. The real-time moving speed when moving away and the relative distance when moving away are input to obtain the relative position risk assessment impact factor corresponding to the moving speed when moving away and the relative position risk assessment impact factor corresponding to the relative distance when moving away from the mapping set. The above impact factors are all extracted from the mine car database.

[0073] It should be understood that, from the above embodiment, when the target person is gradually approaching the mine car, the smaller the relative position, indicating that the closer he is to the mine car, the faster the movement speed, and the larger the relative position risk assessment index, indicating that the potential risk level of the target person's position is greater; when the target person is gradually moving away from the mine car, the faster the movement speed, the larger the relative position, indicating that the farther he is from the mine car, and the smaller the relative position risk assessment index, indicating that the potential risk level of the target person's position is smaller.

[0074] It should be explained that the locations of the target personnel and the mine car can be obtained by carrying a global positioning system receiver, the movement speed of the target personnel can be obtained by calculating the ratio of the target personnel's moving distance in a preset time period to the time period, and the moving distance of the target personnel can be obtained by calculating the difference between the positions at the start time point and the end time point.

[0075] In a specific embodiment, by continuously monitoring the action trajectory data of each target person within a preset time period and calculating the change in distance relative to the mine car in real time, dynamic tracking of personnel behavior can be achieved. This real-time performance helps to timely discover and warn of possible risk behaviors, such as people gradually approaching the mine car, especially the behavior of intending to kick the mine car. By comparing the relative distance difference between the previous adjacent instantaneous recording point and the current time point, it is possible to accurately judge whether the target person is approaching or away from the mine car. Combined with the preset critical action speed and critical relative distance, the risk status of each target person relative to the mine car can be more accurately assessed, thereby identifying which personnel may be implementing or preparing to implement dangerous behaviors of kicking the mine car. Once it is identified that the target person has a tendency to approach the mine car and may implement kicking behavior, the system can immediately issue an early warning signal to remind the supervisory and management personnel to pay attention and take necessary measures, such as increasing monitoring, conducting safety education or implementing physical isolation, so as to effectively prevent the occurrence of dangerous behaviors such as kicking the mine car.

[0076] According to the overlap of abnormal behavior characteristics of each target personnel and the relative position risk assessment index, a comprehensive analysis is conducted to obtain the abnormal behavior risk assessment index of each target personnel. Based on the abnormal behavior risk assessment index of each target personnel, a risk assessment is conducted on the behavior of personnel leaning on the mine car, and the assessment results are fed back.

[0077] Specifically, a comprehensive analysis is performed to obtain the abnormal behavior risk assessment index of each target person. The specific analysis process is: based on the overlap of abnormal behavior characteristics of each target person and the relative position risk assessment index, a comprehensive analysis is performed to obtain the abnormal behavior risk assessment index of each target person. The abnormal behavior risk assessment index of each target person is used to quantitatively assess the risk level of abnormal behavior of each target person.

[0078] In a specific embodiment, the numerical expression of the abnormal behavior risk assessment index of each target person is:

[0079] ;

[0080] In the formula, represents the abnormal behavior risk assessment index of the nth target person, Indicates the overlap degree of abnormal behavior characteristics of the nth target person, Represents the relative position risk assessment index of the nth target person, Indicates the abnormal behavior risk assessment impact factor corresponding to the preset abnormal behavior feature overlap, Indicates the abnormal behavior risk assessment impact factor corresponding to the preset relative position risk assessment index.

[0081] It should be explained that, in the above embodiment, the abnormal behavior risk assessment influence factor corresponding to the feature overlap and the abnormal behavior risk assessment influence factor corresponding to the relative position risk assessment index are both set to 1. When the target person's abnormal behavior feature overlap is greater and the relative position risk assessment index is greater, the target person's abnormal behavior risk assessment index is also greater, which indicates that the risk level of the target person's abnormal behavior is greater.

[0082] It should be understood that if Figure 2 As shown, curve a represents the relationship between the target person's abnormal behavior feature overlap and the abnormal behavior risk assessment index when the target person's relative position risk assessment index is 0.64, curve b represents the relationship between the target person's abnormal behavior feature overlap and the abnormal behavior risk assessment index when the target person's relative position risk assessment index is 0.89, and curve c represents the relationship between the target person's abnormal behavior feature overlap and the abnormal behavior risk assessment index when the target person's relative position risk assessment index is 1.03.

[0083] It should be explained that, in this embodiment, the value range of the abnormal behavior risk assessment impact factor corresponding to the feature overlap and the abnormal behavior risk assessment impact factor corresponding to the relative position risk assessment index is between 0 and 1. Through the relationship between the feature overlap and the relative position risk assessment index and the abnormal behavior risk assessment index in the historical data, a mapping set of the feature overlap and the relative position risk assessment index and their corresponding abnormal behavior risk assessment impact factors is established, and the real-time feature overlap and relative position risk assessment index are input to obtain the abnormal behavior risk assessment impact factor corresponding to the feature overlap and the abnormal behavior risk assessment impact factor corresponding to the relative position risk assessment index from the mapping set. The above impact factors are all extracted from the mine car database.

[0084] It should be understood that the above embodiment can capture the abnormal behavior characteristics of the target person more comprehensively by combining the overlap of abnormal behavior characteristics and the relative position risk assessment index, and conducting a comprehensive analysis from the two dimensions of behavior and position. By comparing the behavioral characteristics of the target person with the preset abnormal behavior characteristic image, the characteristics similar to the behavior of the lying and pedaling mine car can be accurately identified, reducing misjudgment and missed judgment. The system can process the behavioral data and position data of the target person in real time and calculate the abnormal behavior risk assessment index in time. When the abnormal behavior risk assessment index of the target person exceeds the preset threshold, the system can immediately issue an early warning signal to remind the management personnel to take corresponding measures. In this embodiment, not only the behavioral characteristics of the target person are considered, but also the position information of the target person relative to the mine car is combined to make the risk assessment more comprehensive and accurate. The risk assessment results obtained based on big data analysis and intelligent algorithms provide strong support for the decision-making of managers. According to the risk assessment results, managers can adjust management strategies, strengthen supervision and education of high-risk personnel, and reduce the probability of dangerous behaviors such as lying and pedaling mine cars.

[0085] In a specific embodiment, the data in the mine car database includes the minimum safe distance between personnel and the mine car, each reference abnormal behavior characteristic image, and the relative position risk assessment influencing factor corresponding to the action speed and the relative position risk assessment influencing factor corresponding to the relative distance, etc. The data in the mine car database can be collected through various sensors installed on the mine car, such as global positioning system locators, accelerometers, temperature sensors, etc., and can also be obtained by using specially designed software and hardware systems to manage and monitor the operating status of the mine car and then analyze it.

[0086] Furthermore, a risk assessment of personnel's behavior of lying on the mine car is performed according to the abnormal behavior risk assessment index of each target person. The specific assessment process is: according to the abnormal behavior risk assessment index of each target person, the abnormal behavior risk assessment index threshold of the target person is extracted from the mine car database, and the abnormal behavior risk assessment index of each target person is compared with the abnormal behavior risk assessment index threshold of the target person. If the abnormal behavior risk assessment index of a target person is higher than the abnormal behavior risk assessment index threshold of the target person, it indicates that the risk of abnormal behavior of the target person is relatively high, and the target person is marked as a high-risk person. If the abnormal behavior risk assessment index of a target person is equal to the abnormal behavior risk assessment index threshold of the target person, it indicates that the risk of abnormal behavior of the target person is at a medium level, and the target person is marked as a medium-risk person. If the abnormal behavior risk assessment index of a target person is lower than the abnormal behavior risk assessment index threshold of the target person, it indicates that the risk of abnormal behavior of the target person is relatively low, and the target person is marked as a low-risk person, and the assessment result is fed back.

[0087] In a specific embodiment, the specific feedback process of the above-mentioned evaluation results is to record the risk level of each target person and associate it with the personal information of the target person for subsequent tracking and processing. Once a person is marked as a high-risk or medium-risk person, the target person and other relevant personnel should be notified immediately by SMS, email or live broadcast. For high-risk personnel, in addition to immediate notification, personalized safety training plans or intervention measures should be provided to help the target person change unsafe behavior habits. The evaluation results will be generated into a detailed evaluation report, including but not limited to: a list of people marked as high-risk, medium-risk and low-risk; statistics on the number of people at each risk level; specific behavioral descriptions of high-risk and medium-risk personnel and the risks they may cause; recommended safety measures and improvement suggestions.

[0088] It should be understood that the above embodiment divides the target personnel into high risk, medium risk and low risk levels by setting different risk assessment index thresholds, which makes it easier for managers to take corresponding management measures according to the risk level, thereby improving the pertinence and efficiency of management. According to the risk level, it is also possible to reasonably allocate resources to ensure that high-risk areas and high-risk personnel receive more attention and resource investment. This is conducive to reducing the risk of abnormal behavior of target personnel and improving the safety and efficiency of mining operations.

[0089] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for identifying the behavior of a person leaning on a mine car, characterized in that: include: The mine car operation area is delineated to obtain the target area, and the personnel in the target area are marked as target personnel; The behavioral action data of the target personnel are collected and processed to obtain the overlap of abnormal behavioral characteristics of each target personnel; The specific process of collecting the behavior data of the target person is as follows: Collecting real-time video image data of a target person within a preset time period, including the target person's movement trajectory data and behavior action data within the preset time period, wherein the target person's behavior action data includes the target person's behavior feature image at each instantaneous recording point; The above processing results in the coincidence degree of abnormal behavior characteristics of each target person. The specific processing process is as follows: According to the behavior feature images of each target person at each instantaneous recording point, several key joint position points are deployed on the image, and a two-dimensional coordinate system is constructed to mark each key joint position point, which is recorded as , where n represents the number of each target person, , m represents the total number of target personnel, i represents the number of each key joint, , s represents the total number of key joints, r represents the number of each instantaneous recording point, , h represents the total number of instantaneous recording points, Represents the horizontal coordinate of the i-th key joint position point of the n-th target person at the r-th instantaneous recording point, It represents the ordinate of the i-th key joint position point of the n-th target person at the r-th instantaneous recording point, Represents the i-th key joint position point of the n-th target person at the r-th instantaneous recording point. The reference abnormal behavior feature images are extracted from the mine car database. According to the behavior feature images of each target person at each instantaneous recording point, similar abnormal behavior feature images are matched and the key joint position points of similar abnormal behavior feature images are collected, which are recorded as ,in, Represents the horizontal coordinate of the i-th key joint position point of the reference standard, Represents the ordinate of the i-th key joint position point of the reference standard, represents the i-th key joint position point, and the overlap degree of abnormal behavior characteristics of each target person is obtained by comprehensive analysis; The target personnel's action trajectory data is collected and processed to obtain the relative position risk assessment index of each target personnel; According to the overlap of abnormal behavior characteristics of each target person and the relative position risk assessment index, a comprehensive analysis is performed to obtain the abnormal behavior risk assessment index of each target person, and the risk assessment of the abnormal behavior of the personnel is conducted according to the abnormal behavior risk assessment index of each target person, and the assessment results are fed back; The abnormal behavior risk assessment index of each target person is used to quantitatively assess the risk level of abnormal behavior of each target person; The specific numerical expression of the abnormal behavior risk assessment index of each target person is: ； In the formula, represents the abnormal behavior risk assessment index of the nth target person, It indicates the overlap degree of abnormal behavior characteristics of the nth target person. Represents the relative position risk assessment index of the nth target person, Indicates the abnormal behavior risk assessment impact factor corresponding to the preset abnormal behavior feature overlap, Indicates the abnormal behavior risk assessment impact factor corresponding to the preset relative position risk assessment index.

2. According to the method for identifying the behavior of a person leaning on a mine car according to claim 1, it is characterized by: The target area is obtained by demarcating the mine car operation area, and the specific process is as follows: The minimum safe distance between personnel and the mine car is extracted from the mine car database, and an area is delineated with the mine car location point as the center and the minimum safe distance between personnel and the mine car as the radius. The delineated area is marked as the target area, and the personnel in the target area are marked as target personnel.

3. According to claim 1, a method for identifying the behavior of a person leaning on a mine car is characterized by: The target personnel's movement trajectory data includes the distance between each target personnel and the mine car and the movement speed of each target personnel within a preset time period.

4. According to claim 1, a method for identifying the behavior of a person leaning on a mine car is characterized by: The relative position risk assessment index of each target person is obtained through processing, and the specific process is as follows: The distance between each target person and the mine car is marked as the relative distance of each target person, and the relative distance between the previous adjacent instantaneous recording point and the current time point is extracted from the action trajectory data of each target person within the preset time period, and the relative distance between the previous adjacent instantaneous recording point and the current time point is subtracted. If the difference is greater than 0, it indicates that the target person is gradually approaching the mine car, and if the difference is less than 0, it indicates that the target person is gradually moving away from the mine car; According to the movement speed and relative distance of each target person within a preset time period, the critical movement speed and critical relative distance are extracted from the mine car database, and the relative position risk assessment index of each target person is obtained through comprehensive analysis.

5. According to claim 1, a method for identifying the behavior of a person leaning on a mine car is characterized by: The risk assessment of abnormal behavior of personnel is performed according to the abnormal behavior risk assessment index of each target person. The specific assessment process is as follows: According to the abnormal behavior risk assessment index of each target person, the threshold of the abnormal behavior risk assessment index of the target person is extracted from the mine car database, and the abnormal behavior risk assessment index of each target person is compared with the threshold of the abnormal behavior risk assessment index of the target person. If the abnormal behavior risk assessment index of a target person is higher than the threshold of the abnormal behavior risk assessment index of the target person, it indicates that the risk of abnormal behavior of the target person is relatively high, and the target person is marked as a high-risk person. If the abnormal behavior risk assessment index of a target person is equal to the threshold of the abnormal behavior risk assessment index of the target person, it indicates that the risk of abnormal behavior of the target person is at a medium level, and the target person is marked as a medium-risk person. If the abnormal behavior risk assessment index of a target person is lower than the threshold of the abnormal behavior risk assessment index of the target person, it indicates that the risk of abnormal behavior of the target person is relatively low, and the target person is marked as a low-risk person, and the assessment result is fed back.

6. According to claim 4, a method for identifying the behavior of a person leaning on a mine car is characterized by: The relative position risk assessment index of each target person is specifically expressed as: ； In the formula, It represents the relative position risk assessment index of the nth target person, and n represents the number of each target person. , m represents the total number of target personnel, Indicates the action speed of the nth target person, Indicates the relative distance of the nth target person, Indicates the critical action speed, represents the critical relative distance, Indicates the relative distance of the nth target person at the previous adjacent instantaneous recording point, Indicates the relative distance of the nth target person at the current time point, Indicates the relative position risk assessment impact factor corresponding to the preset approach action speed, Indicates the relative position risk assessment impact factor corresponding to the preset relative distance when approaching. Indicates the relative position risk assessment impact factor corresponding to the preset moving speed when moving away. Indicates the relative position risk assessment impact factor corresponding to the preset relative distance.

Images