Mine safety production behavior monitoring and data analysis system

Abstract

The invention relates to the technical field of mine safety monitoring, solves the technical problems that miner behavior monitoring is insufficient, equipment fault diagnosis is not accurate enough, and dangerous condition judgment is not intelligent, and particularly relates to a mine safety production behavior monitoring and data analysis system. Acquiring miner position information, miner behavior image information and equipment operation state data in a mine; the behavior analysis module is used for analyzing the miner behavior image information based on an image recognition algorithm to obtain miner behavior safety indexes; the fault judgment module is used for judging whether the equipment breaks down or not by using a fault diagnosis algorithm; the danger judgment module judges whether a dangerous condition that the miner gets close to the fault equipment exists or not, and a dangerous condition index is obtained; and the safety evaluation module comprehensively evaluates mine safety production behaviors. The intelligent level of miner behavior monitoring is improved, the accuracy of equipment fault diagnosis is enhanced, and danger association analysis between miners and equipment is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of mine safety monitoring, and in particular to a mine safety production behavior monitoring and data analysis system. Background Art

[0002] With the rapid development of industrial automation and information technology, safety issues in mine production operations have attracted widespread attention. Mining operations are characterized by harsh environments and high operational risks, making miner safety an increasingly pressing issue. Traditional mine safety management relies primarily on manual inspections and equipment monitoring to ensure safe mine production. However, manual inspections are inefficient and susceptible to human factors, while equipment monitoring lacks comprehensive monitoring and timely response to miner behavior. In recent years, with the rapid development of technologies such as artificial intelligence, the Internet of Things, big data, and machine learning, a growing number of new technologies have been applied to mine safety management, particularly intelligent monitoring and analysis of miner behavior, equipment status, and the operating environment. Through data collection, behavioral analysis, and equipment status monitoring, miner behavior and equipment operating status can be understood in real time, effectively reducing safety risks in mine operations.

[0003] However, existing mine safety monitoring systems mostly focus on detecting equipment failures and lack real-time monitoring and analysis of miners' behavior. Miners' safe behavior and operating standards are important guarantees for safe mine production, but traditional technologies cannot effectively evaluate miners' safe operating behavior. Existing mine equipment monitoring technologies mostly rely on traditional condition monitoring methods for the prediction and diagnosis of equipment failures. They have difficulty in dealing with equipment failure judgment under complex working conditions and cannot provide timely warnings for potential safety hazards. In traditional mine safety management, the judgment of the dangerous correlation between miners' behavior and equipment failures is relatively simple. It is impossible to analyze in real time whether miners are close to faulty equipment and take preventive measures in time, resulting in difficulty in obtaining a quick response once an accident occurs. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides a mine safety production behavior monitoring and data analysis system, which solves the technical problems of insufficient miner behavior monitoring, inaccurate equipment fault diagnosis, and lack of intelligent judgment of dangerous conditions. It improves the intelligence level of miner behavior monitoring, enhances the accuracy of equipment fault diagnosis, and realizes the risk correlation analysis between miners and equipment.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: a mine safety production behavior monitoring and data analysis system, the system comprising:

[0006] Data acquisition module, used to obtain miners' location information, miners' behavior image information and equipment operation status data in the mine;

[0007] A behavior analysis module is used to perform in-depth feature extraction and analysis on the miner behavior image information based on the image recognition algorithm to obtain a miner behavior safety index;

[0008] The fault diagnosis module is used to pre-process the equipment operation status data to obtain the equipment operation parameters and use the fault diagnosis algorithm to determine whether the equipment has a fault;

[0009] A danger judgment module, which judges whether there is a dangerous situation where the miner is close to the faulty equipment based on the miner's location information and the equipment fault judgment result, and obtains a danger situation index;

[0010] The safety assessment module constructs a safety assessment model based on the dangerous condition index and the miner behavior safety index based on a machine learning algorithm to comprehensively evaluate the mine safety production behavior.

[0011] Furthermore, the process of obtaining the location information of miners in the mine is as follows:

[0012] Install a UWB positioning base station and equip each miner with a UWB positioning tag. The tag corresponds to the miner one by one and is used to send and receive wireless signals.

[0013] Calculate the distance between the miner and each base station based on the signal's propagation speed in the air and the signal's propagation time from the tag to the base station;

[0014] Based on the distance between the miner and each base station, the three-dimensional coordinate position of the miner in the mine is calculated using the trilateral positioning algorithm. The specific process is as follows:

[0015] According to the principle of trilateral positioning, the distance between the miner and the base station is expressed by the following formula:

[0016]

[0017] Where d1 represents the distance from base station 1 to the miner, d2 represents the distance from base station 2 to the miner, and d3 represents the distance from base station 3 to the miner. The position of base station 1 is (x1, y1, z1), the position of base station 2 is (x2, y2, z2), the position of base station 3 is (x3, y3, z3), and the coordinate position of the miner is (x, y, z).

[0018] Subtracting the three formulas from each other to eliminate the square terms results in a system of linear equations. Solving the system of linear equations yields the values of x, y, and z, as well as the miner's three-dimensional coordinates (x, y, z).

[0019] Furthermore, the distance between the miner and each base station is calculated based on the propagation speed of the signal in the air and the propagation time of the signal from the tag to the base station. The specific process is as follows:

[0020] Obtain the additional delay time and signal propagation time caused by multipath effects;

[0021] The propagation velocity is corrected by the propagation velocity coefficient to obtain the effective propagation velocity;

[0022] The distance between the miner and each base station is calculated based on the effective propagation speed, the additional delay time caused by the multipath effect, and the signal propagation time. The calculation formula is:

[0023]

[0024] Where, d i is the distance between the miner and the i-th base station, c is the propagation speed of the signal in the air; a is the propagation speed coefficient, t i is the signal propagation time, Δt i It is the additional delay time caused by multipath effect.

[0025] Furthermore, the process of performing deep feature extraction and analysis on the miner behavior image information based on the image recognition algorithm to obtain the miner behavior safety index is as follows:

[0026] The acquired miner behavior image information is preprocessed, and deep feature extraction and analysis is performed on the miner image information based on the convolutional neural network model;

[0027] Extract local features from the image and fuse them through pooling and fully connected layers. Then, train the convolutional neural network model and optimize its parameters to determine whether the miners are wearing helmets.

[0028] Classify the image based on the trained model to determine whether the miner is wearing a hard hat;

[0029] If the image recognition result shows that the miner is not wearing a safety helmet, the miner's behavioral safety index is determined to be 0;

[0030] If it is determined that the miner is wearing a safety helmet, the wearing features of the safety helmet in the image are further analyzed to calculate the miner's behavioral safety index.

[0031] Furthermore, if it is determined that the miner is wearing a helmet, the wearing features of the helmet in the image are further analyzed to calculate the miner's behavioral safety index as follows:

[0032] Use an object detection algorithm to identify the bounding boxes of the hardhat and the miner’s head in the image;

[0033] Calculate the center coordinates of the helmet and the miner's head based on the detected helmet bounding box coordinates and the miner's head bounding box coordinates;

[0034] The degree of position difference is calculated based on the coordinates of the center point of the helmet and the miner's head. The calculation formula is:

[0035]

[0036] Where D is the degree of position difference, (x a ,y a ) is the coordinate of the center point of the miner's head, (x h ,y h ) is the coordinate of the center point of the helmet;

[0037] Use image segmentation algorithm to segment the hatband from the image and extract the shape, color and texture features of the hatband;

[0038] The shape, color, and texture features of the hatband are used to form a feature vector, and the decision function f(x) in the support vector machine model is used to classify the feature vector x to determine whether the hatband is buckled.

[0039] If the decision function f(x)>0, the hat strap is judged to be buckled, and the degree of difference in the buckled condition is 1;

[0040] If the decision function f(x)≤0, it is judged that the hat strap is not fastened, and the degree of difference in the hat strap fastening situation is 0.

[0041] Use the Canny edge detection algorithm to extract the contours of the helmet and the miner's head respectively, and calculate the contour areas of the helmet and the miner's head;

[0042] Overlap the outlines of the helmet and the miner's head and calculate the area of the overlapping region;

[0043] and calculate the minimum distance between the hard hat outline and the miner's head outline;

[0044] Evaluate the fit between the helmet and the head by comprehensively considering the overlapping area and the contour distance;

[0045] The difference in the position of the helmet and the miner's head, the difference in the buckling of the hat strap, and the degree of fit are comprehensively analyzed to calculate the miner's behavioral safety index.

[0046] Furthermore, the calculation formula for the degree of fit between the helmet and the head is:

[0047]

[0048] Where F is the fit between the helmet and the head, A c The area of the overlapping region, A hat is the outline area of the helmet, A head is the miner's head contour area, α is the silhouette coefficient, dmin is the minimum distance between the hard hat outline and the miner's head outline, and β is the distance coefficient.

[0049] Furthermore, the comprehensive analysis of the difference in the position of the helmet and the miner's head, the difference in the buckle of the chin strap, and the degree of fit is used to calculate the miner's behavior safety index. The specific calculation formula is:

[0050]

[0051] Where F is the fit between the helmet and the head, φ1 is the fit coefficient between the helmet and the head, M is the difference in the buckle condition of the chin strap, φ2 is the difference coefficient of the buckle condition of the chin strap, D is the position difference, φ3 is the position difference coefficient, and S is the behavioral safety index.

[0052] Furthermore, the process of determining whether a device failure occurs is as follows:

[0053] Establish a mathematical model for the equipment and determine whether the equipment has failed by comparing the difference between the equipment operating parameters and the model prediction data. If the difference exceeds a preset threshold, it is determined that the equipment may have failed.

[0054] Set a normal parameter range for each device operating parameter, compare the acquired device operating parameters with the preset normal parameter range, and mark the parameter as abnormal if it exceeds the normal parameter range;

[0055] If the difference exceeds the preset threshold and the device operating parameters are abnormal, the device will fail;

[0056] If the difference does not exceed the preset threshold and the device operating parameters are not abnormal, the device is operating normally;

[0057] If the difference exceeds the preset threshold and the equipment operating parameters are normal, or if the difference does not exceed the preset threshold and the equipment operating parameters are abnormal, the equipment will be included in the fault alarm list and the equipment operating status data will continue to be tested.

[0058] Furthermore, the process of obtaining the danger status index is as follows:

[0059] According to the equipment fault judgment results, the faulty equipment is selected and the coordinate position of the faulty equipment is recorded as (a, b, c). Combined with the coordinate position of the miner (x, y, z), the distance between the faulty equipment and the miner is calculated. The specific calculation formula is:

[0060]

[0061] Where, d m is the distance between the faulty equipment and the miner;

[0062] For each miner m, calculate the distance d between miner m and each faulty device in turn m ;

[0063] Compare the distance d between miner m and each faulty device m with the danger distance threshold Y;

[0064] If d m ≥Y, the miner is in a dangerous situation of approaching the faulty device; and record the number of faulty devices when each miner m is in a dangerous situation;

[0065] If d m <Y, the miner is not in a dangerous situation of approaching the faulty device; sum up the number of faulty devices when each miner m is in a dangerous situation to obtain the danger situation index G.

[0066] Furthermore, the process of comprehensively evaluating the safe production behavior of the mine is as follows:

[0067] Based on the danger situation index and the miner behavior safety index, construct a safety evaluation model using a machine learning algorithm. The expression of the safety evaluation model is:

[0068]

[0069] In the formula, Q is the safety evaluation coefficient, ρ1 is the weight of the miner behavior safety index, S is the weight of the miner behavior safety index, ρ2 is the weight of the danger situation index, and G is the danger situation index;

[0070] Set two thresholds T low and T high , where T low <T high ;

[0071] If the safety evaluation coefficient Q is less than the threshold T low , it means that the overall safe production behavior of the mine is good, and there are few dangerous situations and unsafe behaviors of personnel;

[0072] If the safety evaluation coefficient Q is greater than the threshold T low and less than the threshold T high , it indicates that there are certain potential safety hazards, but the overall situation is controllable;

[0073] If the safety evaluation coefficient Q is greater than the threshold T high , it means that the mine is in a high safety risk state, and immediate measures need to be taken to eliminate dangerous situations and correct unsafe behaviors of personnel.

[0074] By means of the above technical solutions, the present invention provides a mine safety production behavior monitoring and data analysis system, which at least has the following beneficial effects:

[0075] 1. This invention can monitor various safety behaviors during mine production in real time. Using sensors, cameras, and other equipment to collect on-site data, it automatically identifies dangerous behaviors within the mine and issues timely safety warnings through an intelligent early warning mechanism to prevent accidents. The system can conduct in-depth analysis of the collected data, identify potential safety hazards and behavioral patterns, and provide detailed statistical reports. These data analysis results provide decision support for mine managers, helping to optimize safety production processes and improve work efficiency.

[0076] 2. Through real-time monitoring and feedback of employee behavior, this system can encourage employees to pay more attention to their own safety practices. Visual data feedback can also strengthen employees' safety behavior awareness and reduce illegal operations and unsafe behaviors. The system can record each employee's safety behavior data, providing companies with a traceable record of their behavior. This accumulated data can help companies improve employee safety training, optimize safety management processes, and provide a basis for future reviews and accident responsibility attribution.

[0077] 3. Leveraging intelligent monitoring technology, this system can effectively reduce the incidence of safety accidents caused by human negligence and illegal operations, ensuring smooth mine production and improving the overall safety level of the mine. The system's high degree of automation enables round-the-clock, comprehensive monitoring of safety activities within the mine, reducing the shortcomings of manual monitoring, improving work efficiency, and saving labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0079] Figure 1 This is a flow chart of a mine safety production behavior monitoring and data analysis system of the present invention. DETAILED DESCRIPTION

[0080] To make the above-mentioned objectives, features, and advantages of the present invention more clearly understood, the present invention is further described below in detail with reference to the accompanying drawings and specific embodiments. This will enable a full understanding of how this application uses technical means to solve technical problems and achieve technical effects, and to implement the invention accordingly.

[0081] This embodiment uses an image recognition algorithm to conduct an in-depth analysis of the miners' behavior to obtain a more accurate miners' safety behavior index, thereby evaluating whether the miners have safety operation risks. The equipment operation status data is pre-processed using a fault diagnosis algorithm to obtain more accurate equipment operation parameters, determine the occurrence of equipment failures, and issue early warnings. Combining the miners' real-time location information with the equipment failure results, it is analyzed whether the miners are in the dangerous area of the equipment failure, and a more intelligent judgment of dangerous conditions is achieved. Through a machine learning algorithm, a comprehensive mine safety assessment model is established based on the miners' behavior safety index and the dangerous condition index, which evaluates the mine's production safety behavior as a whole and provides a decision-making basis for mine safety management. This embodiment proposes a mine production safety behavior monitoring and data analysis system, which improves the intelligence level of miners' behavior monitoring, enhances the accuracy of equipment fault diagnosis, and realizes the analysis of dangerous associations between miners and equipment. As Figure 1 As shown in the figure, the system consists of a data acquisition module, a behavior analysis module, a fault judgment module, a danger judgment module and a safety assessment module.

[0082] As a further implementation example, the data acquisition module is used to obtain the location information of miners in the mine, the image information of miners' behaviors, and the operating status data of equipment. The process of obtaining the location information of miners in the mine is as follows:

[0083] Multiple UWB positioning base stations are installed at suitable locations within the mine, covering areas where miners are likely to be. Each miner is equipped with a UWB positioning tag, which is unique to the miner and is used to send and receive wireless signals.

[0084] The positioning tag sends UWB pulse signals to the surrounding space at a certain frequency. After receiving these signals, the positioning base station records the timestamp of the signal arrival. At the same time, the positioning base station also sends corresponding signals to the positioning tag so that the tag can obtain relevant information about the base station.

[0085] The distance between the miner and each base station is calculated based on the signal's propagation speed in the air and the signal's propagation time from the tag to the base station. The specific process is as follows:

[0086] Since signals may reflect or refract during propagation, the path that the signal takes to reach the base station is not a straight line, but multiple different paths. This may increase the propagation time. Therefore, the influence of multipath effect is considered to obtain the additional delay time caused by multipath effect and the signal propagation time.

[0087] Since the propagation speed is affected by environmental factors, the propagation speed is corrected by the propagation speed coefficient to obtain the effective propagation speed;

[0088] The distance between the miner and each base station is calculated based on the effective propagation speed, the additional delay time caused by the multipath effect, and the signal propagation time. The specific calculation formula is:

[0089]

[0090] Where, d i is the distance between the miner and the i-th base station, c is the propagation speed of the signal in the air; a is the propagation speed coefficient, t i is the signal propagation time, Δt i It is the additional delay time caused by multipath effect.

[0091] Based on the distance between the miner and each base station, the miner's three-dimensional coordinate position in the mine is calculated using a trilateral positioning algorithm. The specific process is as follows:

[0092] According to the principle of trilateral positioning, the distance between the miner and the base station can be expressed by the following formula:

[0093]

[0094] Where d1 represents the distance from base station 1 to the miner, d2 represents the distance from base station 2 to the miner, and d3 represents the distance from base station 3 to the miner. The position of base station 1 is (x1, y1, z1), the position of base station 2 is (x2, y2, z2), the position of base station 3 is (x3, y3, z3), and the coordinate position of the miner is (x, y, z).

[0095] Subtracting the three formulas from each other and eliminating the square terms yields two linear equations. Through these two linear equations, we can solve the values of x, y, and z to obtain the miner's three-dimensional coordinates (x, y, z).

[0096] The calculated miner location information is transmitted to the server of the ground monitoring center through the wireless communication network, and the real-time location of the miner in the mine is displayed in a visual manner on the software interface of the monitoring center.

[0097] The process of obtaining miner behavior image information is as follows:

[0098] Explosion-proof surveillance cameras are installed in the main operating areas of the mine to ensure coverage of the miners' main activities. These cameras need to have high-definition and low-light characteristics to adapt to the complex environment of the mine;

[0099] The surveillance camera captures the scene inside the mine in real time, converts the optical signal into an electrical signal, and then compresses and encodes it into digital image data through a video encoding algorithm;

[0100] The collected image data is transmitted to the ground storage server via the industrial Ethernet in the mine to obtain image information of miners' behavior. To facilitate subsequent query and analysis, the data is usually stored according to certain time intervals and storage strategies, such as daily or weekly data archiving.

[0101] The process of obtaining device operating status data is as follows:

[0102] Sensors are installed at key locations on the equipment to sense the equipment's operating status and output electrical signals. The signals output by the sensors are collected and converted into numbers that can be processed by computers to obtain equipment operating status data.

[0103] A miner behavior judgment module performs in-depth feature extraction and analysis on the miner behavior image information based on an image recognition algorithm to obtain a miner behavior safety index;

[0104] The acquired images of miners' behaviors are preprocessed and deep feature extraction and analysis is performed on the images based on a convolutional neural network model. Multiple convolutional layers are used to extract local features from the images, and feature fusion is performed through pooling and fully connected layers. The model is trained and its parameters are optimized to improve image recognition accuracy and accurately determine whether miners are wearing helmets.

[0105] The trained model will classify each image to determine whether the miner is wearing a hard hat;

[0106] If the image recognition result shows that the miner is not wearing a helmet, the miner's behavioral safety index is directly determined to be 0;

[0107] If it is determined that the miner is wearing a hard hat, the wearing characteristics of the hard hat in the image are further analyzed, including the position of the hard hat, the buckle condition of the chin strap, and the fit with the head, to calculate the miner's behavioral safety index.

[0108] The miner's behavioral safety index is calculated as follows:

[0109] Use the target detection algorithm to identify the bounding boxes of the helmet and the miner's head in the image. Based on the detected bounding box coordinates of the helmet and the miner's head, calculate the center point coordinates of the helmet and the miner's head. The position difference degree is calculated by the center point coordinates of the helmet and the miner's head. The calculation formula is:

[0110]

[0111] Where D is the degree of position difference, (x a ,y a ) is the coordinate of the center point of the miner's head, (x h ,yh ) are the coordinates of the center point of the helmet.

[0112] Use image segmentation algorithm to segment the hatband from the image, extract the shape, color and texture features of the hatband, and form a feature vector based on the shape, color and texture features of the hatband;

[0113] Use the decision function f(x) in the support vector machine model to classify the feature vector x and determine whether the hatband is buckled; the process is as follows:

[0114] Principal component analysis (PCA) is used to reduce the dimensionality of the feature vector to remove the correlation between features and reduce the amount of calculation;

[0115] Use the ReliefF algorithm to select the features after dimensionality reduction, find the features that contribute most to classification, and improve the accuracy and efficiency of classification;

[0116] The processed feature vectors and corresponding labels (buttoned is 1, unbuttoned is 0) are divided into training set and test set;

[0117] Use the training set to train the support vector machine model, select the radial basis kernel function, and optimize the model parameters through grid search and cross-validation methods;

[0118] The calculation formula of the decision function f(x) is:

[0119]

[0120] Where a i is the Lagrange multiplier, y i is the label of the training sample, b is the bias term, n s represents the number of support vectors, K(x i , x) is the radial basis kernel function;

[0121] The calculation formula of the radial basis kernel function is as follows:

[0122] K(x i ,x)=exp(-r||x i , x|| 2 )

[0123] In the formula, K(x i , x) is the radial basis kernel function, r is the kernel coefficient, ||x i , x|| are two samples x i The Euclidean distance from x;

[0124] The button judgement parameter is calculated based on the decision function f(x), and the formula is:

[0125]

[0126] Where S is the buckle judgment parameter, f(x train ) is the decision function value of all samples in the training set;

[0127] The value range of S is [0,1]. The closer the S value is to 1, the higher the confidence of the classification;

[0128] If the decision function f(x)>0, and the buckle judgment parameter is greater than the set threshold, then the hat strap is judged to be buckled, and the degree of difference in the hat strap buckle situation is 1;

[0129] If the decision function f(x)≤0, and the buckle judgment parameter is greater than the set threshold, it is judged that the hat strap is not buckled, and the difference degree of the hat strap buckle situation is 0.

[0130] It should be noted that when the button judgment parameter is less than the set threshold, x is reclassified.

[0131] Use the Canny edge detection algorithm to extract the contours of the helmet and the miner's head respectively, and calculate the contour areas of the helmet and the miner's head;

[0132] Overlap the outlines of the helmet and the miner's head and calculate the area of the overlapping region;

[0133] and calculate the minimum distance between the hard hat outline and the miner's head outline;

[0134] The overlap area and contour distance are comprehensively considered to evaluate the fit between the helmet and the head. The fit between the helmet and the head is calculated as follows:

[0135]

[0136] Where F is the fit between the helmet and the head, A c The area of the overlapping region, A hat is the outline area of the helmet, A head is the miner's head contour area, α is the silhouette coefficient, d min is the minimum distance between the hard hat outline and the miner's head outline, and β is the distance coefficient.

[0137] The miner's behavior safety index is calculated by comprehensively analyzing the difference in the position of the helmet and the miner's head, the difference in the buckle of the hat strap, and the degree of fit. The specific calculation formula is:

[0138]

[0139] Where F is the fit between the helmet and the head, is the coefficient of fit between the helmet and the head, M is the difference in the buckle condition of the chin strap, is the coefficient of difference in the buckle condition of the chin strap, D is the degree of position difference, is the coefficient of position difference, and S is the behavioral safety index.

[0140] The equipment fault judgment module pre-processes the equipment operation status data to obtain the equipment operation parameters, and uses the fault diagnosis algorithm to determine whether the equipment has a fault;

[0141] Preprocess the equipment operation status data and analyze the preprocessed data to obtain equipment operation parameters;

[0142] Use the model-based fault diagnosis algorithm to determine whether the equipment has a fault. The process is as follows:

[0143] Establish a mathematical model for the equipment and determine whether the equipment has failed by comparing the actual operating data with the model's predicted data. If the difference exceeds a preset threshold, the equipment is considered to be faulty.

[0144] According to the equipment's design requirements and historical operating data, a normal parameter range is set for each equipment operating parameter. The acquired equipment operating parameters are compared with the preset normal parameter range. If the equipment operating parameter exceeds the normal parameter range, the parameter is marked as abnormal.

[0145] If the difference exceeds the preset threshold and the equipment operating parameters are outside the normal parameter range, the equipment will fail;

[0146] If the difference does not exceed the preset threshold and the device operating parameters do not exceed the normal parameter range, the device is operating normally;

[0147] If the difference exceeds the preset threshold and the equipment operating parameters do not exceed the normal parameter range, or if the difference does not exceed the preset threshold and the equipment operating parameters exceed the normal parameter range, the equipment will be included in the fault alarm list and the equipment operating status data will continue to be detected.

[0148] A danger situation judgment module, which judges whether there is a danger situation caused by the miner being close to the faulty equipment through a spatial position association algorithm based on the miner's location information and the equipment fault judgment result, and obtains a danger situation index;

[0149] According to the equipment fault judgment results, the faulty equipment is selected and the coordinate position of the faulty equipment is recorded as (a, b, c). Combined with the coordinate position of the miner (x, y, z), the distance between the faulty equipment and the miner is calculated. The specific calculation formula is:

[0150]

[0151] Where, d m is the distance between the faulty equipment and the miner;

[0152] According to the type of equipment, the nature of the fault, and the possible scope of harm, set corresponding dangerous distance thresholds Y for different types of faulty equipment. For electrical equipment faults, the possible set dangerous distance threshold is 5 meters; for ventilation equipment faults, the possible set dangerous distance threshold is 3 meters.

[0153] For each miner m, calculate the distance between miner m and each faulty equipment in turn, and compare the distance between miner m and each faulty equipment with the dangerous distance threshold. If d m ≥Y, it means that the miner is in a dangerous situation of approaching the faulty equipment; if d m <Y, it means that the miner is not in a dangerous situation of approaching the faulty equipment; and record the number of faulty equipment when each miner m is in a dangerous situation.

[0154] Sum up the number of faulty equipment when each miner m is in a dangerous situation to obtain the danger situation index G.

[0155] The safety production behavior monitoring module constructs a safety assessment model based on a machine learning algorithm according to the danger situation index combined with the miner behavior safety index, and comprehensively evaluates the safety production behavior of the mine.

[0156] Construct a safety assessment model based on a machine learning algorithm according to the danger situation index combined with the miner behavior safety index, and generate a safety assessment coefficient Q. The expression of the safety assessment model is:

[0157]

[0158] In the formula, Q is the safety assessment coefficient, ρ1 is the weight of the miner behavior safety index, S is the weight of the miner behavior safety index, ρ2 is the weight of the danger situation index, and G is the danger situation index.

[0159] Set two thresholds T low and T high , where T low <T high ;

[0160] If the safety assessment coefficient Q is less than the threshold T low , it indicates that the overall safety production behavior of the mine is good, and the dangerous situations and unsafe behaviors of personnel are very few;

[0161] If the safety assessment coefficient Q is greater than the threshold T low and less than the threshold T high , it means that there are certain potential safety hazards, but the overall situation is controllable;

[0162] If the safety assessment coefficient Q is greater than the threshold T high , it indicates that the mine is in a high safety risk state, and immediate measures need to be taken to eliminate dangerous situations and correct the unsafe behaviors of personnel.

[0163] Those skilled in the art will appreciate that all or part of the steps in the above-mentioned embodiment methods can be accomplished by instructing the relevant hardware through a program. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0164] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to in detail. For the above embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For relevant parts, please refer to the partial description of the method embodiments.

[0165] The above embodiments provide a detailed introduction to the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A mine safety production behavior monitoring and data analysis system, characterized in that: The system includes: Data acquisition module, used to obtain miners' location information, miners' behavior image information and equipment operation status data in the mine; A behavior analysis module, which performs in-depth feature extraction and analysis on the miner behavior image information based on an image recognition algorithm to obtain a miner behavior safety index; The fault diagnosis module is used to pre-process the equipment operation status data to obtain the equipment operation parameters and use the fault diagnosis algorithm to determine whether the equipment has a fault; A danger judgment module, which judges whether there is a dangerous situation where the miner is close to the faulty equipment based on the miner's location information and the equipment fault judgment result, and obtains a danger situation index; The safety assessment module constructs a safety assessment model based on the dangerous condition index and the miner behavior safety index based on the machine learning algorithm to comprehensively evaluate the mine safety production behavior.

2. A mine safety production behavior monitoring and data analysis system according to claim 1, characterized in that: The process of obtaining the location information of miners in the mine is as follows: Install a UWB positioning base station and equip each miner with a UWB positioning tag. The tag corresponds to the miner one-to-one and is used to send and receive wireless signals. Calculate the distance between the miner and each base station based on the signal's propagation speed in the air and the signal's propagation time from the tag to the base station; Based on the distance between the miner and each base station, the three-dimensional coordinate position of the miner in the mine is calculated using the trilateral positioning algorithm. The specific process is as follows: According to the principle of trilateral positioning, the distance between the miner and the base station is expressed by the following formula: Where d1 represents the distance from base station 1 to the miner, d2 represents the distance from base station 2 to the miner, and d3 represents the distance from base station 3 to the miner. The position of base station 1 is (x1, y1, z1), the position of base station 2 is (x2, y2, z2), the position of base station 3 is (x3, y3, z3), and the coordinate position of the miner is (x, y, z). Subtracting the three formulas from each other to eliminate the square terms results in a system of linear equations. Solving the system of linear equations yields the values of x, y, and z, as well as the miner's three-dimensional coordinates (x, y, z).

3. A mine safety production behavior monitoring and data analysis system according to claim 2, characterized in that: The distance between the miner and each base station is calculated based on the signal's propagation speed in the air and the signal's propagation time from the tag to the base station. The specific process is as follows: Obtain the additional delay time and signal propagation time caused by multipath effects; The propagation velocity is corrected by the propagation velocity coefficient to obtain the effective propagation velocity; The distance between the miner and each base station is calculated based on the effective propagation speed, the additional delay time caused by the multipath effect, and the signal propagation time. The calculation formula is: Where, d i is the distance between the miner and the i-th base station, c is the propagation speed of the signal in the air; a is the propagation speed coefficient, t i is the signal propagation time, Δt i It is the additional delay time caused by multipath effect.

4. A mine safety production behavior monitoring and data analysis system according to claim 3, characterized in that: The process of performing deep feature extraction and analysis on the miner behavior image information based on the image recognition algorithm to obtain the miner behavior safety index is as follows: The acquired miner behavior image information is preprocessed, and deep feature extraction and analysis is performed on the miner image information based on the convolutional neural network model; Extract local features from the image and fuse them through pooling and fully connected layers. Then, train the convolutional neural network model and optimize its parameters to determine whether the miners are wearing helmets. Classify the image based on the trained model to determine whether the miner is wearing a hard hat; If the image recognition result shows that the miner is not wearing a safety helmet, the miner's behavioral safety index is determined to be 0; If it is determined that the miner is wearing a safety helmet, the wearing features of the safety helmet in the image are further analyzed to calculate the miner's behavioral safety index.

5. A mine safety production behavior monitoring and data analysis system according to claim 4, characterized in that: If it is determined that the miner is wearing a helmet, the wearing features of the helmet in the image are further analyzed to calculate the miner's behavioral safety index as follows: Use an object detection algorithm to identify the bounding boxes of the hardhat and the miner’s head in the image; Calculate the center coordinates of the helmet and the miner's head based on the detected helmet bounding box coordinates and the miner's head bounding box coordinates; The degree of position difference is calculated based on the coordinates of the center point of the helmet and the miner's head. The calculation formula is: Where D is the degree of position difference, (x a ,y a ) is the coordinate of the center point of the miner's head, (x h ,y h ) is the coordinate of the center point of the helmet; Use image segmentation algorithm to segment the hatband from the image and extract the shape, color and texture features of the hatband; The feature vector is formed according to the shape, color and texture features of the hatband. Use the decision function f(x) in the support vector machine model to classify the feature vector x and determine whether the hatband is buckled. When the button judgment parameter is less than the set threshold, reclassify x; Use the Canny edge detection algorithm to extract the contours of the helmet and the miner's head respectively, and calculate the contour areas of the helmet and the miner's head; Overlap the outlines of the helmet and the miner's head and calculate the area of the overlapping region; and calculate the minimum distance between the hard hat outline and the miner's head outline; The overlap area and contour distance are comprehensively considered to evaluate the fit between the helmet and the head. The formula is: Where F is the fit between the helmet and the head, A c The area of the overlapping region, A hat is the outline area of the helmet, A head is the miner's head contour area, α is the silhouette coefficient, d min is the minimum distance between the hard hat outline and the miner's head outline, and β is the distance coefficient; The difference in the position of the helmet and the miner's head, the difference in the buckling of the hat strap, and the degree of fit are comprehensively analyzed to calculate the miner's behavioral safety index.

6. A mine safety production behavior monitoring and data analysis system according to claim 5, characterized in that: The feature vector is composed according to the shape, color and texture features of the hatband, and the feature vector x is classified using the decision function f(x) in the support vector machine model. The process is as follows: The principal component analysis method is used to reduce the dimension of the feature vector; Divide the processed feature vectors and corresponding labels into training sets and test sets; The support vector machine model is trained using the training set. By selecting the radial basis kernel function, the decision function f(x) in the support vector machine model is used to classify the feature vector x, and the model parameters are optimized through grid search and cross-validation methods. The calculation formula of the decision function f(x) of the support vector machine model is: Where a i is the Lagrange multiplier, y i is the label of the training sample, b is the bias term, n s represents the number of support vectors, K(x i , x) is the radial basis kernel function; The calculation formula of the radial basis kernel function is as follows: K(x i ,x)=exp(-r||x i ,x|| 2 ) In the formula, K(x i , x) is the radial basis kernel function, r is the kernel coefficient, ||x i , x|| are two samples x i The Euclidean distance from x; The button judgement parameter is calculated based on the decision function f(x), and the formula is: Where S is the buckle judgment parameter, f(x train ) is the decision function value of all samples in the training set; If the decision function f(x)>0, and the buckle judgment parameter is greater than the set threshold, then the hat strap is judged to be buckled, and the degree of difference in the hat strap buckle situation is 1; If the decision function f(x)≤0, and the buckle judgment parameter is greater than the set threshold, it is judged that the hat strap is not buckled, and the difference degree of the hat strap buckle situation is 0.

7. A mine safety production behavior monitoring and data analysis system according to claim 5, characterized in that: The comprehensive analysis is to calculate the miner's behavior safety index by analyzing the difference between the position of the helmet and the miner's head, the difference between the buckle of the helmet strap and the degree of fit. The specific calculation formula is: Where F is the fit between the helmet and the head, φ1 is the fit coefficient between the helmet and the head, M is the difference in the buckle condition of the chin strap, φ2 is the difference coefficient of the buckle condition of the chin strap, D is the position difference, φ3 is the position difference coefficient, and S is the behavioral safety index.

8. A mine safety production behavior monitoring and data analysis system according to claim 1, characterized in that: The process of determining whether a device failure occurs is as follows: Establish a mathematical model for the equipment and determine whether the equipment has failed by comparing the difference between the equipment operating parameters and the model prediction data. If the difference exceeds a preset threshold, it is determined that the equipment may have failed. Set a normal parameter range for each device operating parameter, compare the acquired device operating parameters with the preset normal parameter range, and mark the parameter as abnormal if it exceeds the normal parameter range; If the difference exceeds the preset threshold and the device operating parameters are abnormal, the device will fail; If the difference does not exceed the preset threshold and the device operating parameters are not abnormal, the device is operating normally; If the difference exceeds the preset threshold and the equipment operating parameters are normal, or if the difference does not exceed the preset threshold and the equipment operating parameters are abnormal, the equipment will be included in the fault alarm list and the equipment operating status data will continue to be tested.

9. A mine safety production behavior monitoring and data analysis system according to claim 8, characterized in that: The process of obtaining the dangerous situation index is as follows: According to the equipment fault judgment results, the faulty equipment is selected and the coordinate position of the faulty equipment is recorded as (a, b, c). Combined with the coordinate position of the miner (x, y, z), the distance between the faulty equipment and the miner is calculated. The specific calculation formula is: Where, d m is the distance between the faulty equipment and the miner; For each miner m, calculate the distance d between miner m and each faulty device in turn m ; The distance d between miner m and each faulty device m Compare with the danger distance threshold Y; If d m ≥Y, the miner is in a dangerous situation close to the faulty equipment; and the number of faulty equipment when each miner m is in a dangerous situation is recorded; If d m < is less than Y, there is no dangerous situation for the miner near the faulty equipment; the number of faulty equipment when each miner m is in a dangerous situation is summed to obtain the danger situation index G.

10. A mine safety production behavior monitoring and data analysis system according to claim 9, characterized in that: The process of comprehensive assessment of mine safety production behavior is as follows: According to the dangerous situation index and the miner behavior safety index, a safety assessment model is constructed based on the machine learning algorithm. The safety assessment model expression is: Where Q is the safety assessment coefficient, ρ1 is the miner behavior safety index weight, S is the miner behavior safety index weight, ρ2 is the dangerous situation index weight, and G is the dangerous situation index; Set two thresholds T low and T high , where T low <T high ; If the safety assessment factor Q is less than the threshold T low , indicating that the mine's overall safety production behavior is good, with few dangerous conditions and unsafe behaviors of personnel; If the safety assessment factor Q is greater than the threshold T low Less than the threshold T high , it means there are certain safety risks, but the overall situation is controllable; If the safety assessment factor Q is greater than the threshold T high , indicating that the mine is in a high safety risk state and immediate measures need to be taken to eliminate dangerous conditions and correct unsafe behaviors of personnel.

Images