Coal mine safety monitoring data fusion processing method and system

Abstract

The invention discloses a coal mine safety monitoring data fusion processing method and system, and relates to the technical field of data fusion processing, and the method comprises the following steps: S1, screening a roof effective index, S2, determining a characteristic value, S3, outputting an index risk level, S4, screening a to-be-grown point, S5, analyzing tunnel deformation, and S6, executing an early warning prompt operation. According to the method, a plurality of basic indexes are constructed according to a plurality of roof accident analysis cases, the basic indexes are further processed, and after multi-level indexes are determined, effective indexes are screened out according to scoring results under different dimensions, so that on one hand, the requirements of roof accident risk assessment can be met to the greatest extent, and on the other hand, the risk assessment efficiency is improved; on the other hand, interference of redundant data is reduced, it is guaranteed that the risk analysis result is more accurate, meanwhile, the simulation three-dimensional curved surface of the coal mine roadway is constructed through the method, once the roadway deforms, the specific position of a deformation occurrence point can be accurately analyzed, and an operator can take measures in time for intervention.

Description

Technical Field

[0001] The present invention relates to the technical field of data fusion processing, and in particular to a method and system for fusion processing of coal mine safety monitoring data. Background Art

[0002] The coal mine safety monitoring data fusion processing method mainly integrates the monitored data for analysis and processing. Once an abnormal situation occurs, the method will issue an early warning to help coal mine managers take timely measures to intervene. The invention patent with application number 202411195382.8 discloses "a coal mine safety monitoring data fusion processing system and method based on the Internet of Things, involving the field of electronic digital data processing technology. First, dividing the coal mine monitoring area can improve the accuracy of safety detection. The risks of each sub-area can be independently evaluated and controlled, making the response to local problems more timely and effective. Then, the coal mine safety monitoring data of each coal mine monitoring sub-area in the coal mine safety monitoring platform is analyzed to improve the accuracy of safety assessment. By analyzing the equipment data of each coal mine monitoring sub-area, equipment failures or abnormalities can be accurately identified, and abnormal results of each abnormal coal mine monitoring sub-area can be obtained. Early warning prompts and adjustments are carried out, which helps to quickly respond to and deal with potential risks, thereby enhancing the overall safety protection level. "

[0003] The above-mentioned existing technology solves the problems of being unable to accurately identify the spatial position and time dimension of the monitoring sites. However, when the system is running, it does not screen the monitoring indicators, resulting in redundant data that may interfere with the determination of the risk level. At the same time, it does not conduct a more detailed analysis of the weights of all indicators. Once the errors in some indicator data are large, it will directly lead to deviations in multi-level indicator data, thereby affecting the accuracy of the coal mine roof risk analysis. In addition, the system cannot analyze the degree of deformation of the coal mine tunnels, and the operators cannot take corresponding measures to intervene. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for fusion processing of coal mine safety monitoring data to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for fusion processing of coal mine safety monitoring data, comprising the following steps:

[0006] S1. Screening effective roof indicators: Use multiple roof accident analysis cases to construct several basic indicators, perform cluster analysis on all basic indicators, determine the secondary indicators and primary indicators corresponding to each basic indicator, use ten experts to score each indicator according to four dimensions, and construct an initial indicator scoring matrix based on these data. Combine the initial indicator scoring matrix with the set weights to obtain a weighted indicator scoring matrix, determine the maximum and minimum scores corresponding to each indicator, and mark invalid indicators and valid indicators based on these scores;

[0007] S2, determine the eigenvalue: after obtaining the data corresponding to each indicator in multiple mining areas, construct the corresponding covariance matrix according to these data, and determine all the eigenvalues ​​and eigenvectors generated by the matrix, count the proportion of each eigenvalue, and add them up one by one. When the accumulated result is greater than the first threshold, stop the operation and output all the eigenvalues ​​involved in the previous round of accumulation. Determine the number of important indicators according to the number of eigenvalues ​​involved in the accumulation, use the component evaluation algorithm to analyze the component values ​​corresponding to the overall indicators and the effective indicators, use these values ​​to further determine the number of effective indicators, and delete invalid indicators and corresponding data;

[0008] S3. Output indicator risk level: After setting a relative coefficient for each indicator, the corresponding decision matrix is ​​calculated through the relative coefficient of the indicator, and the subjective weight of each indicator is analyzed using the decision matrix. After obtaining the score value corresponding to each indicator data, the correlation coefficient corresponding to each indicator is determined, and the objective weight corresponding to the indicator is calculated according to the correlation coefficient. After determining the weight coefficient, the corresponding combined weight is calculated according to these data, and the risk level is divided into four levels. The corresponding characteristic coefficient is set for each risk level. The characteristic coefficient corresponding to each indicator is analyzed according to the basic indicator data collected in the specified mining area and the combined weight of the indicator. The similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level is calculated. After determining the risk level of all indicators according to the similarity, it is output through a visual interface;

[0009] S4, screening the points to be grown: extracting the light streak image data of the coal mine tunnel, determining the corresponding target area in the image data, performing a grayscale binarization operation on the area, determining the foreground point and the background point according to the binarized image data, constructing an inscribed circle in the non-background point area with each foreground point as the center point, and storing the center point in a key point set, calculating the distance value between each pixel point in the set and the nearest background point, using the value to replace the original pixel value, and screening the points to be grown according to these data;

[0010] S5. Analyze the deformation of the tunnel: Analyze the refined target area according to the position of the growth point, count all the center points and specific positions in the area, calculate the actual coordinates according to the pixel coordinates corresponding to each center point using a coordinate conversion algorithm, and use 3D software to splice these data to obtain the simulated 3D surface corresponding to the coal mine tunnel, extract the initial 3D surface of the coal mine tunnel, slice and compare the initial 3D surface and the simulated 3D surface, and analyze the specific position and volume of the deformation point of the tunnel;

[0011] S6. Execute early warning prompt operations: If there are indicators with a risk level of four in the coal mine roof monitoring and the tunnel deformation level is severe deformation, it will be reported directly to the control center and a prompt message will be issued using voice broadcasting. If the number of first-level indicators with a risk level of three in the coal mine roof monitoring is greater than one, it will be reported directly to the control center. If the number of first-level indicators with a risk level of three in the coal mine roof monitoring is less than or equal to one and the tunnel deformation level is slight deformation, a prompt will be issued using voice broadcasting.

[0012] Preferably, the step S1 specifically includes the following steps:

[0013] S101, extracting multiple roof accident analysis cases, transferring them to a data analysis tool, and using the analysis tool to conceptualize and categorize the sentences in the cases to obtain multiple roof-related basic indicators, wherein the data analysis tool is specifically NVivo analysis software;

[0014] S102. Perform cluster analysis on all basic indicators to determine the secondary indicators corresponding to each basic indicator. Perform cluster analysis on all secondary indicators again to determine the primary indicators corresponding to each secondary indicator. The primary indicators include human factors, technical factors, environmental factors and management factors. The secondary indicators include protection concepts, safety behaviors, support technologies, geological structures, coal seam characteristics, roof stability and safety inspections. The basic indicators include safety awareness, protection capabilities, risk identification skills, inspection execution, empty roof operations, support strength, support methods, support density, fault drop, fault inclination, roof fold degree, goaf hanging roof area, coal seam depth, coal seam inclination, coal seam hardness, roof lithology, roof integrity, safety inspection, operation rectification, acceptance management, emergency rescue execution and safety signs.

[0015] S103. Utilize ten experts to score each indicator according to the four dimensions of relevance, independence, operability and reliability, calculate the average value of each indicator in different dimensions based on multiple groups of scores, and construct an initial indicator scoring matrix;

[0016] S104, setting weights corresponding to different dimensions, combining the initial indicator scoring matrix with the weights, and obtaining a weighted indicator scoring matrix, determining the maximum and minimum scores corresponding to each indicator, and calculating these data using a distance analysis algorithm to obtain the distance between each indicator and its corresponding maximum and minimum scores. If the distance between the indicator and the maximum score is less than or equal to the distance between the indicator and the minimum score, it is marked as an invalid indicator, otherwise it is marked as a valid indicator. The distance analysis algorithm is specifically as follows:

[0017]

[0018] in, Indicates the distance between index i and the maximum score, S i ′ represents the distance between index i and the minimum score, t ij Indicates the score of the i-th indicator corresponding to the j-th dimension, Indicates the maximum score corresponding to each indicator, t′ j It represents the minimum score corresponding to each indicator, i and j represent parameters, and m represents the total number of dimensions.

[0019] Preferably, the step S2 specifically includes the following steps:

[0020] S201, after obtaining data corresponding to various indicators in multiple mining areas, standardize these data, calculate the correlation coefficients between different indicators based on the processed data, construct the corresponding covariance matrix using the correlation coefficients, and determine all eigenvalues ​​and eigenvectors corresponding to the matrix;

[0021] S202, add up all the eigenvalues ​​to calculate the total eigenvalue, count the proportion of each eigenvalue, arrange all the eigenvalues ​​in descending order, accumulate the proportions of the eigenvalues ​​one by one, and when the accumulated result is greater than the first threshold, stop the operation and output all the eigenvalues ​​involved in the accumulation in the previous comparison.

[0022] Preferably, the step S2 specifically further comprises the following steps:

[0023] S203, determine the number of important indicators according to the number of eigenvalues ​​involved in the accumulation, use these eigenvalues ​​and eigenvectors to calculate the data corresponding to each important indicator, extract the score values ​​obtained by ten experts based on the indicator data, standardize these data, and use the component evaluation algorithm to calculate the processed data to obtain the corresponding component value. The component evaluation algorithm is specifically:

[0024]

[0025] Among them, Z abIndicates the score proportion corresponding to the bth data in the ath important indicator, t ab represents the score value corresponding to the bth data in the ath important indicator, S b represents the entropy value corresponding to the b-th data, C represents the component value, ln(ρ) represents the maximum value of all entropies of important indicators, a and b represent parameters, h represents the number of important indicators, and k represents the number of entropy values;

[0026] S204. After counting the component values ​​corresponding to the overall indicator data, the component values ​​corresponding to the effective indicators are calculated in the same way, and the ratio between the effective indicators and the overall indicators is analyzed. If the ratio is lower than the second threshold, the effective indicators are re-screened, otherwise the invalid indicators and the corresponding data are deleted.

[0027] Preferably, the step S3 specifically includes the following steps:

[0028] S301. After setting the relative coefficient for each indicator, compare each indicator one by one. If the relative coefficient of the current indicator is higher than that of other indicators, set the corresponding judgment coefficient to one. If the relative coefficients are equal, set the corresponding judgment coefficient to zero. If the relative coefficient of the current indicator is lower than that of other indicators, set the corresponding judgment coefficient to negative one. Construct a judgment matrix based on all the judgment coefficients of each indicator, use the judgment matrix to deduce the corresponding decision matrix, and analyze the subjective weight of each indicator through the decision matrix.

[0029] S302, after obtaining the score values ​​corresponding to each indicator data, construct an initial matrix based on these data, normalize them, determine the correlation coefficient corresponding to each indicator, and calculate the objective weight corresponding to the indicator based on the correlation coefficient;

[0030] S303, after receiving the subjective weight and objective weight of each indicator, construct multiple conditional functions based on these data, conduct joint analysis on the conditional functions to obtain weight coefficients, and calculate the corresponding combined weight according to the weight coefficients and the subjective weight and objective weight of each indicator;

[0031] S304, dividing the risk level into four levels, from level 1 to level 4, namely, negligible risk, acceptable risk, acceptable risk after repair, and unacceptable risk, and setting a corresponding characteristic coefficient for each risk level;

[0032] S305. After analyzing multiple risk parameters corresponding to each indicator based on the basic indicator data collected in the designated mining area, determine the characteristic coefficient corresponding to each risk parameter, combine it with the combined weight of the indicator, and thus obtain the characteristic coefficient of each secondary indicator. Follow the same steps to determine the characteristic coefficient of the primary indicator, calculate the similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level, take the risk level with the highest similarity as the actual risk level of the indicator, and after determining the risk levels of all indicators, output them through a visual interface.

[0033] Preferably, the step S4 specifically includes the following steps:

[0034] S401, using a line scanning laser to scan the coal mine tunnel, and using a camera to shoot the light stripes generated by the scanning, and transmitting the obtained multiple image data to a database, wherein the line scanning laser is specifically an LDLY1670 laser, and the camera is specifically a HY-1091 measurement camera;

[0035] S402, after extracting the light streak image data, distinguish all the pixels in the image according to the set color component range, store the pixels that meet the set range into a target set, and after determining the area formed by all the pixels in the set, use the area as the target area;

[0036] S403, filter the target area using a 3×3 size template, perform grayscale binarization on the processed image data, determine the foreground point and the background point according to the binarized image data, and construct an inscribed circle in the non-background point area with each foreground point as the center point. If the current inscribed circle belongs to a subset of other inscribed circles, no operation is performed, otherwise the corresponding center point is stored in the key point set;

[0037] S404, after traversing each pixel in the set, calculate the distance value between the pixel and the nearest background point, and use the value to replace the original pixel value;

[0038] S405, after setting the size of the local neighborhood, compare the value of each pixel in the set with the values ​​of other foreground points in the area, and if the distance value of the current pixel in any direction is the largest, determine the pixel as the point to be grown;

[0039] S406, after counting all the points to be grown in the target area, determine the number of foreground points in the neighborhood around the point to be grown, if the number is lower than the quantity threshold, set the pixel value corresponding to the point to be grown to zero, otherwise do not perform any operation.

[0040] Preferably, the step S5 specifically includes the following steps:

[0041] S501, analyzing the refined target area according to the position of the point to be grown, traversing each pixel point in the area row by row, determining the position of the pixel point whose gray value is greater than the preset value in each row, calculating the position of the center point corresponding to the row according to the gray value and position of these pixel points, and counting all the center points and specific positions in the area and storing them in the database;

[0042] S502, after obtaining the focal length and scale factor of the camera, using a coordinate conversion algorithm to calculate the actual coordinates according to the pixel coordinates corresponding to each center point;

[0043] S503, after determining the center point and specific position of all image data, use 3D software to splice these data to obtain a complete point cloud map, and perform surface fitting on it to determine the simulated 3D surface corresponding to the coal mine tunnel, wherein the 3D software is specifically CloudCompare 3D software;

[0044] S504, extract the initial three-dimensional surface of the coal mine roadway, slice the initial three-dimensional surface and the simulated three-dimensional surface to obtain multiple initial point cloud slices and simulated point cloud slices, compare the initial point cloud slices and the simulated point cloud slices according to the slice position, analyze the specific position and volume of the roadway deformation point, use deep learning technology to analyze these data, and determine the deformation level of the current roadway.

[0045] The coal mine safety monitoring data fusion processing system comprises an index generation unit, an effective index determination unit, a risk level output unit, a to-be-grown point screening unit, a deformation point analysis unit and an early warning prompt unit;

[0046] The indicator generation unit constructs multiple basic indicators using multiple roof accident analysis cases, performs cluster analysis on all basic indicators, determines the secondary indicators and primary indicators corresponding to each basic indicator, uses ten experts to score each indicator according to four dimensions, and marks invalid indicators and valid indicators according to these scores;

[0047] After the effective index determination unit obtains data corresponding to each index in multiple mining areas, it determines the number of important indexes based on these data, uses a component evaluation algorithm to analyze the component values ​​corresponding to the overall index and the effective index, uses these values ​​to further determine the number of effective indexes, and deletes invalid indicators and corresponding data;

[0048] After the risk level output unit determines the weight coefficient, subjective weight and objective weight, the corresponding combined weight is calculated according to these data, the risk level is divided into four levels, and a corresponding characteristic coefficient is set for each risk level. The characteristic coefficient corresponding to each indicator is analyzed according to the basic indicator data collected in the designated mining area and the combined weight of the indicator, and the similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level is calculated. After the risk level of all indicators is determined according to the similarity, it is output through a visual interface;

[0049] The to-be-grown point screening unit extracts the light streak image data of the coal mine tunnel, determines the corresponding target area in the image data, calculates the distance value between each pixel point in the area and the nearest background point, uses the value to replace the original pixel value, and screens out the to-be-grown points based on these data;

[0050] The deformation point analysis unit analyzes the refined target area according to the position of the growth point, counts all the center points and specific positions in the area, and then uses three-dimensional software to splice these data to obtain the simulated three-dimensional surface corresponding to the coal mine roadway, and analyzes the specific position and volume of the roadway deformation occurrence point;

[0051] If the early warning prompt unit shows indicators of risk level four in the coal mine roof monitoring and the tunnel deformation level is severe deformation, it will report directly to the control center and send out prompt information through voice broadcast. If the number of first-level indicators with risk level three in the coal mine roof monitoring is greater than one, it will report directly to the control center. If the number of first-level indicators with risk level three in the coal mine roof monitoring is less than or equal to one and the tunnel deformation level is slight deformation, voice broadcast will be used for prompting.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. The present invention constructs several basic indicators based on multiple roof accident analysis cases, further processes these basic indicators, determines the multi-level indicators, and then screens out effective indicators according to the scoring results under different dimensions. Such a design can meet the needs of roof accident risk assessment to the greatest extent on the one hand, and reduce the interference of redundant data on the other hand, ensuring that the risk analysis results are more accurate. In addition, the method obtains data corresponding to various indicators in multiple mining areas, and repeatedly verifies whether the currently screened effective indicators can cover information from all aspects of the roof monitoring process based on these data. If the effective indicators cannot cover the information contained in all indicators, they will be screened again, otherwise the invalid indicators will be deleted to avoid redundant data occupying storage space during the analysis process, greatly shortening the time required for data calculation;

[0054] 2. The present invention calculates the subjective weight and objective weight of each indicator, combines them with the weight coefficient, determines the combined weight of each indicator, and analyzes the corresponding primary and secondary indicator data based on the combined weight and indicator data, thereby ensuring that the subsequent determination of the risk level of multi-level indicators can be more reliable and accurate. Even if errors occur in some indicator data during the collection process, it will hardly affect the risk level determination of other indicators. At the same time, the method extracts the light streak image data of the coal mine tunnel, analyzes the image data, determines the center point position corresponding to each image data, and constructs a simulated three-dimensional surface based on the position information. Once the tunnel is deformed, it only needs to compare the original three-dimensional surface with the current simulated three-dimensional surface to accurately analyze the specific position of the deformation point, so that the operator can take timely measures to intervene. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 An overall method flow chart is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] See also Figure 1 The present invention provides a technical solution: a method for fusion processing of coal mine safety monitoring data, comprising the following steps:

[0058] S1. Screening effective roof indicators: Use multiple roof accident analysis cases to construct several basic indicators, perform cluster analysis on all basic indicators, determine the secondary indicators and primary indicators corresponding to each basic indicator, use ten experts to score each indicator according to four dimensions, and construct an initial indicator scoring matrix based on these data. Combine the initial indicator scoring matrix with the set weights to obtain a weighted indicator scoring matrix, determine the maximum and minimum scores corresponding to each indicator, and mark invalid indicators and valid indicators based on these scores;

[0059] S2, determine the eigenvalue: after obtaining the data corresponding to each indicator in multiple mining areas, construct the corresponding covariance matrix according to these data, and determine all the eigenvalues ​​and eigenvectors generated by the matrix, count the proportion of each eigenvalue, and add them up one by one. When the accumulated result is greater than the first threshold, stop the operation and output all the eigenvalues ​​involved in the previous round of accumulation. Determine the number of important indicators according to the number of eigenvalues ​​involved in the accumulation, use the component evaluation algorithm to analyze the component values ​​corresponding to the overall indicators and the effective indicators, use these values ​​to further determine the number of effective indicators, and delete invalid indicators and corresponding data;

[0060] S3. Output indicator risk level: After setting a relative coefficient for each indicator, the corresponding decision matrix is ​​calculated through the relative coefficient of the indicator, and the subjective weight of each indicator is analyzed using the decision matrix. After obtaining the score value corresponding to each indicator data, the correlation coefficient corresponding to each indicator is determined, and the objective weight corresponding to the indicator is calculated according to the correlation coefficient. After determining the weight coefficient, the corresponding combined weight is calculated according to these data, and the risk level is divided into four levels. The corresponding characteristic coefficient is set for each risk level. The characteristic coefficient corresponding to each indicator is analyzed according to the basic indicator data collected in the specified mining area and the combined weight of the indicator. The similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level is calculated. After determining the risk level of all indicators according to the similarity, it is output through a visual interface;

[0061] S4, screening the points to be grown: extracting the light streak image data of the coal mine tunnel, determining the corresponding target area in the image data, performing a grayscale binarization operation on the area, determining the foreground point and the background point according to the binarized image data, constructing an inscribed circle in the non-background point area with each foreground point as the center point, and storing the center point in a key point set, calculating the distance value between each pixel point in the set and the nearest background point, using the value to replace the original pixel value, and screening the points to be grown according to these data;

[0062] S5. Analyze the deformation of the tunnel: Analyze the refined target area according to the position of the growth point, count all the center points and specific positions in the area, calculate the actual coordinates according to the pixel coordinates corresponding to each center point using a coordinate conversion algorithm, and use 3D software to splice these data to obtain the simulated 3D surface corresponding to the coal mine tunnel, extract the initial 3D surface of the coal mine tunnel, slice and compare the initial 3D surface and the simulated 3D surface, and analyze the specific position and volume of the deformation point of the tunnel;

[0063] S6. Execute early warning prompt operations: If there are indicators with a risk level of four in the coal mine roof monitoring and the tunnel deformation level is severe deformation, it will be reported directly to the control center and a prompt message will be issued using voice broadcasting. If the number of first-level indicators with a risk level of three in the coal mine roof monitoring is greater than one, it will be reported directly to the control center. If the number of first-level indicators with a risk level of three in the coal mine roof monitoring is less than or equal to one and the tunnel deformation level is slight deformation, a prompt will be issued using voice broadcasting.

[0064] Step S1 specifically includes the following steps:

[0065] S101, extracting multiple roof accident analysis cases, transferring them to a data analysis tool, using the analysis tool to conceptualize and categorize the sentences in the cases, and obtaining multiple roof-related basic indicators, the data analysis tool being specifically NVivo analysis software;

[0066] S102. Perform cluster analysis on all basic indicators to determine the secondary indicators corresponding to each basic indicator. Perform cluster analysis on all secondary indicators again to determine the primary indicators corresponding to each secondary indicator. The primary indicators include human factors, technical factors, environmental factors and management factors. The secondary indicators include protection concepts, safety behaviors, support technologies, geological structures, coal seam characteristics, roof stability and safety inspections. Basic indicators include safety awareness, protection capabilities, risk identification skills, inspection execution, empty roof operations, support strength, support methods, support density, fault drop, fault inclination, roof fold degree, goaf hanging roof area, coal seam depth, coal seam inclination, coal seam hardness, roof lithology, roof integrity, safety inspection, operation rectification, acceptance management, emergency rescue execution and safety signs.

[0067] S103. Utilize ten experts to score each indicator according to the four dimensions of relevance, independence, operability and reliability, calculate the average value of each indicator in different dimensions based on multiple groups of scores, and construct an initial indicator scoring matrix;

[0068] S104. Set the weights corresponding to different dimensions, combine the initial indicator scoring matrix with the weights, and after obtaining the weighted indicator scoring matrix, determine the maximum and minimum scores corresponding to each indicator, and use the distance analysis algorithm to calculate these data to obtain the distance between each indicator and its corresponding maximum and minimum scores. If the distance between the indicator and the maximum score is less than or equal to the distance between the indicator and the minimum score, it is marked as an invalid indicator, otherwise it is marked as a valid indicator. The specific distance analysis algorithm is:

[0069]

[0070] in, Indicates the distance between index i and the maximum score, S i ′ represents the distance between index i and the minimum score, t ij Indicates the score of the i-th indicator corresponding to the j-th dimension, represents the maximum score corresponding to each indicator, j represents the parameter, and m represents the total number of dimensions;

[0071] Step S2 specifically includes the following steps:

[0072] S201, after obtaining data corresponding to various indicators in multiple mining areas, standardize these data, calculate the correlation coefficients between different indicators based on the processed data, construct the corresponding covariance matrix using the correlation coefficients, and determine all eigenvalues ​​and eigenvectors corresponding to the matrix;

[0073] S202, all eigenvalues ​​are added to calculate the total eigenvalue, and after counting the proportion of each eigenvalue, all eigenvalues ​​are arranged in order from large to small, and the proportions of the eigenvalues ​​are accumulated one by one. When the accumulated result is greater than the first threshold, the operation is stopped and all eigenvalues ​​involved in the accumulation in the previous comparison are output;

[0074] Step S2 specifically also includes the following steps:

[0075] S203, determine the number of important indicators according to the number of eigenvalues ​​involved in the accumulation, use these eigenvalues ​​and eigenvectors to calculate the data corresponding to each important indicator, extract the score values ​​obtained by ten experts based on the indicator data, standardize these data, and use the component evaluation algorithm to calculate the processed data to obtain the corresponding component value. The component evaluation algorithm is specifically as follows:

[0076]

[0077] Among them, Z ab Indicates the score proportion corresponding to the bth data in the ath important indicator, t ab represents the score value corresponding to the bth data in the ath important indicator, S b represents the entropy value corresponding to the b-th data, C represents the component value, ln(ρ) represents the maximum value of all entropies of important indicators, a and b represent parameters, h represents the number of important indicators, and k represents the number of entropy values;

[0078] S204, after counting the component values ​​corresponding to the overall index data, calculate the component values ​​corresponding to the effective index in the same way, and analyze the ratio between the effective index and the overall index. If the ratio is lower than the second threshold, re-screen the effective index, otherwise delete the invalid index and the corresponding data;

[0079] Step S3 specifically includes the following steps:

[0080] S301. After setting the relative coefficient for each indicator, compare each indicator one by one. If the relative coefficient of the current indicator is higher than that of other indicators, set the corresponding judgment coefficient to one. If the relative coefficients are equal, set the corresponding judgment coefficient to zero. If the relative coefficient of the current indicator is lower than that of other indicators, set the corresponding judgment coefficient to negative one. Construct a judgment matrix based on all the judgment coefficients of each indicator, use the judgment matrix to deduce the corresponding decision matrix, and analyze the subjective weight of each indicator through the decision matrix.

[0081] S302, after obtaining the score values ​​corresponding to each indicator data, construct an initial matrix based on these data, normalize them, determine the correlation coefficient corresponding to each indicator, and calculate the objective weight corresponding to the indicator based on the correlation coefficient;

[0082] S303, after receiving the subjective weight and objective weight of each indicator, multiple conditional functions are constructed based on these data, and the conditional functions are jointly analyzed to obtain weight coefficients, and the corresponding combined weights are calculated according to the weight coefficients and the subjective weight and objective weight of each indicator. The conditional function is specifically:

[0083]

[0084] Among them, D min represents the distance function value, w represents the combined weight vector, w a represents the subjective weight vector, w b represents the objective weight vector, r1 represents the subjective weight coefficient, and r2 represents the objective weight coefficient;

[0085] S304, dividing the risk level into four levels, from level 1 to level 4, namely, negligible risk, acceptable risk, acceptable risk after repair, and unacceptable risk, and setting a corresponding characteristic coefficient for each risk level;

[0086] S305. After analyzing multiple risk parameters corresponding to each indicator based on the basic indicator data collected in the designated mining area, determine the characteristic coefficient corresponding to each risk parameter, combine it with the combined weight of the indicator, and thus obtain the characteristic coefficient of each secondary indicator. According to the same steps, determine the characteristic coefficient of the primary indicator, calculate the similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level, take the risk level with the highest similarity as the actual risk level of the indicator, and after determining the risk levels of all indicators, output them through a visual interface;

[0087] Step S4 specifically includes the following steps:

[0088] S401, using a line scanning laser to scan the coal mine tunnel, and using a camera to shoot the light stripes generated by the scanning, and transmitting the obtained multiple image data to a database, wherein the line scanning laser is specifically an LDLY1670 laser, and the camera is specifically a HY-1091 measurement camera;

[0089] S402, after extracting the light streak image data, distinguish all the pixels in the image according to the set color component range, store the pixels that meet the set range into a target set, and after determining the area formed by all the pixels in the set, use the area as the target area;

[0090] S403, filter the target area using a 3×3 size template, perform grayscale binarization on the processed image data, determine the foreground point and the background point according to the binarized image data, and construct an inscribed circle in the non-background point area with each foreground point as the center point. If the current inscribed circle belongs to a subset of other inscribed circles, no operation is performed, otherwise the corresponding center point is stored in the key point set;

[0091] S404, after traversing each pixel in the set, calculate the distance value between the pixel and the nearest background point, and use the value to replace the original pixel value;

[0092] S405, after setting the size of the local neighborhood, compare the value of each pixel in the set with the values ​​of other foreground points in the area, and if the distance value of the current pixel in any direction is the largest, determine the pixel as the point to be grown;

[0093] S406, after counting all the points to be grown in the target area, determine the number of foreground points in the neighborhood around the point to be grown, if the number is lower than the quantity threshold, set the pixel value corresponding to the point to be grown to zero, otherwise do not perform any operation;

[0094] Step S5 specifically includes the following steps:

[0095] S501, analyzing the refined target area according to the position of the point to be grown, traversing each pixel point in the area row by row, determining the position of the pixel point whose gray value is greater than the preset value in each row, calculating the position of the center point corresponding to the row according to the gray value and position of these pixel points, and counting all the center points and specific positions in the area and storing them in the database;

[0096] S502, after obtaining the focal length and scale factor of the camera, the actual coordinates are calculated according to the pixel coordinates corresponding to each center point using a coordinate conversion algorithm. The coordinate conversion algorithm is specifically as follows:

[0097]

[0098] Among them, zc represents the scale factor, (u0, v0) represents the pixel coordinates of the center point, fx represents the focal length of the horizontal axis of the camera, and f y Indicates the focal length of the measuring camera in the longitudinal direction, (xw, yw, zw) indicates the actual coordinates of the center point, and a, b, c, d indicates the correlation coefficients in the actual coordinate system;

[0099] S503, after determining the center point and specific position of all image data, use 3D software to stitch these data to obtain a complete point cloud map, and perform surface fitting on it to determine the simulated 3D surface corresponding to the coal mine tunnel, the 3D software is specifically CloudCompare 3D software;

[0100] S504, extracting the initial three-dimensional surface of the coal mine tunnel, slicing the initial three-dimensional surface and the simulated three-dimensional surface to obtain multiple initial point cloud slices and simulated point cloud slices, comparing the initial point cloud slices and the simulated point cloud slices according to the slice positions, analyzing the specific positions and volumes of the tunnel deformation occurrence points, analyzing these data using deep learning technology, and determining the deformation level of the current tunnel;

[0101] A coal mine safety monitoring data fusion processing system, comprising an index generation unit 1, an effective index determination unit 2, a risk level output unit 3, a waiting growth point screening unit 4, a deformation point analysis unit 5 and an early warning prompt unit 6;

[0102] The indicator generation unit 1 constructs multiple basic indicators using multiple roof accident analysis cases, performs cluster analysis on all basic indicators, determines the secondary indicators and primary indicators corresponding to each basic indicator, uses ten experts to score each indicator according to four dimensions, and marks invalid indicators and valid indicators based on these scores;

[0103] After obtaining the data corresponding to each indicator in multiple mining areas, the effective indicator determination unit 2 determines the number of important indicators based on these data, uses the component evaluation algorithm to analyze the component values ​​corresponding to the overall indicator and the effective indicator, uses these values ​​to further determine the number of effective indicators, and deletes invalid indicators and corresponding data;

[0104] After the risk level output unit 3 determines the weight coefficient, subjective weight and objective weight, it calculates the corresponding combined weight according to these data, divides the risk level into four levels, and sets the corresponding characteristic coefficient for each risk level. According to the basic indicator data collected in the specified mining area and the combined weight of the indicator, the characteristic coefficient corresponding to each indicator is analyzed, and the similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level is calculated. After the risk level of all indicators is determined according to the similarity, it is output through a visual interface;

[0105] The waiting growth point screening unit 4 extracts the light streak image data of the coal mine tunnel, determines the corresponding target area in the image data, calculates the distance value between each pixel point in the area and the nearest background point, uses the value to replace the original pixel value, and screens the waiting growth points based on these data;

[0106] The deformation point analysis unit 5 analyzes the refined target area according to the position of the growth point, counts all the center points and specific positions in the area, and then uses three-dimensional software to splice these data to obtain the simulated three-dimensional surface corresponding to the coal mine roadway, and analyzes the specific position and volume of the roadway deformation occurrence point;

[0107] If the early warning prompt unit 6 has indicators of risk level four in the coal mine roof monitoring and the tunnel deformation level is severe deformation, it will be reported directly to the control center and a prompt message will be issued using voice broadcast. If the number of first-level indicators with risk level three in the coal mine roof monitoring is greater than one, it will be reported directly to the control center. If the number of first-level indicators with risk level three in the coal mine roof monitoring is less than or equal to one and the tunnel deformation level is slight deformation, a prompt will be issued using voice broadcast.

[0108] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0109] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A coal mine safety monitoring data fusion processing method, characterized in that: The method comprises the following steps: S1. Screening effective roof indicators: Use multiple roof accident analysis cases to construct several basic indicators, perform cluster analysis on all basic indicators, determine the secondary indicators and primary indicators corresponding to each basic indicator, use ten experts to score each indicator according to four dimensions, and construct an initial indicator scoring matrix based on these data. Combine the initial indicator scoring matrix with the set weights to obtain a weighted indicator scoring matrix, determine the maximum and minimum scores corresponding to each indicator, and mark invalid indicators and valid indicators based on these scores; S2, determine the eigenvalue: after obtaining the data corresponding to each indicator in multiple mining areas, construct the corresponding covariance matrix according to these data, and determine all the eigenvalues ​​and eigenvectors generated by the matrix, count the proportion of each eigenvalue, and add them up one by one. When the accumulated result is greater than the first threshold, stop the operation and output all the eigenvalues ​​involved in the previous round of accumulation. Determine the number of important indicators according to the number of eigenvalues ​​involved in the accumulation, use the component evaluation algorithm to analyze the component values ​​corresponding to the overall indicators and the effective indicators, use these values ​​to further determine the number of effective indicators, and delete invalid indicators and corresponding data; S3. Output indicator risk level: After setting a relative coefficient for each indicator, the corresponding decision matrix is ​​calculated through the relative coefficient of the indicator, and the subjective weight of each indicator is analyzed using the decision matrix. After obtaining the score value corresponding to each indicator data, the correlation coefficient corresponding to each indicator is determined, and the objective weight corresponding to the indicator is calculated according to the correlation coefficient. After determining the weight coefficient, the corresponding combined weight is calculated according to these data, and the risk level is divided into four levels. The corresponding characteristic coefficient is set for each risk level. The characteristic coefficient corresponding to each indicator is analyzed according to the basic indicator data collected in the specified mining area and the combined weight of the indicator. The similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level is calculated. After determining the risk level of all indicators according to the similarity, it is output through a visual interface; S4, screening the points to be grown: extracting the light streak image data of the coal mine tunnel, determining the corresponding target area in the image data, performing a grayscale binarization operation on the area, determining the foreground point and the background point according to the binarized image data, constructing an inscribed circle in the non-background point area with each foreground point as the center point, and storing the center point in a key point set, calculating the distance value between each pixel point in the set and the nearest background point, using the value to replace the original pixel value, and screening the points to be grown according to these data; S5. Analyze the deformation of the tunnel: Analyze the refined target area according to the position of the growth point, count all the center points and specific positions in the area, calculate the actual coordinates according to the pixel coordinates corresponding to each center point using a coordinate conversion algorithm, and use 3D software to splice these data to obtain the simulated 3D surface corresponding to the coal mine tunnel, extract the initial 3D surface of the coal mine tunnel, slice and compare the initial 3D surface and the simulated 3D surface, and analyze the specific position and volume of the deformation point of the tunnel; S6. Execute early warning prompt operations: If there are indicators with a risk level of four in the coal mine roof monitoring and the tunnel deformation level is severe deformation, it will be reported directly to the control center and a prompt message will be issued using voice broadcasting. If the number of first-level indicators with a risk level of three in the coal mine roof monitoring is greater than one, it will be reported directly to the control center. If the number of first-level indicators with a risk level of three in the coal mine roof monitoring is less than or equal to one and the tunnel deformation level is slight deformation, a prompt will be issued using voice broadcasting.

2. A coal mine safety monitoring data fusion processing method according to claim 1, characterized in that: The step S1 specifically includes the following steps: S101, extracting multiple roof accident analysis cases, transferring them to a data analysis tool, and using the analysis tool to conceptualize and categorize the statements in the cases to obtain multiple roof-related basic indicators; S102, performing cluster analysis on all basic indicators to determine the secondary indicator corresponding to each basic indicator, and performing cluster analysis on all secondary indicators again to determine the primary indicator corresponding to each secondary indicator; S103. Utilize ten experts to score each indicator according to the four dimensions of relevance, independence, operability and reliability, calculate the average value of each indicator in different dimensions based on multiple groups of scores, and construct an initial indicator scoring matrix; S104. Set the weights corresponding to different dimensions, combine the initial indicator scoring matrix with the weights, and after obtaining the weighted indicator scoring matrix, determine the maximum and minimum scores corresponding to each indicator, and use the distance analysis algorithm to calculate these data to obtain the distance between each indicator and its corresponding maximum and minimum scores. If the distance between the indicator and the maximum score is less than or equal to the distance between the indicator and the minimum score, it is marked as an invalid indicator, otherwise it is marked as a valid indicator.

3. A coal mine safety monitoring data fusion processing method according to claim 1, characterized in that: The step S2 specifically includes the following steps: S201, after obtaining data corresponding to various indicators in multiple mining areas, standardize these data, calculate the correlation coefficients between different indicators based on the processed data, construct the corresponding covariance matrix using the correlation coefficients, and determine all eigenvalues ​​and eigenvectors corresponding to the matrix; S202, add up all the eigenvalues ​​to calculate the total eigenvalue, count the proportion of each eigenvalue, arrange all the eigenvalues ​​in descending order, accumulate the proportions of the eigenvalues ​​one by one, and when the accumulated result is greater than the first threshold, stop the operation and output all the eigenvalues ​​involved in the accumulation in the previous comparison.

4. A coal mine safety monitoring data fusion processing method according to claim 3, characterized in that: The step S2 specifically further comprises the following steps: S203, determining the number of important indicators according to the number of eigenvalues ​​involved in the accumulation, using these eigenvalues ​​and eigenvectors to calculate the data corresponding to each important indicator, extracting the score values ​​obtained by ten experts based on the indicator data, standardizing these data, and using the component evaluation algorithm to calculate the processed data to obtain the corresponding component value; S204. After counting the component values ​​corresponding to the overall indicator data, the component values ​​corresponding to the effective indicators are calculated in the same way, and the ratio between the effective indicators and the overall indicators is analyzed. If the ratio is lower than the second threshold, the effective indicators are re-screened, otherwise the invalid indicators and the corresponding data are deleted.

5. The method for coal mine safety monitoring data fusion processing according to claim 1, characterized in that: The step S3 specifically comprises the following steps: S301. After setting the relative coefficient for each indicator, compare each indicator one by one. If the relative coefficient of the current indicator is higher than that of other indicators, set the corresponding judgment coefficient to one. If the relative coefficients are equal, set the corresponding judgment coefficient to zero. If the relative coefficient of the current indicator is lower than that of other indicators, set the corresponding judgment coefficient to negative one. Construct a judgment matrix based on all the judgment coefficients of each indicator, use the judgment matrix to deduce the corresponding decision matrix, and analyze the subjective weight of each indicator through the decision matrix. S302, after obtaining the score values ​​corresponding to each indicator data, construct an initial matrix based on these data, normalize them, determine the correlation coefficient corresponding to each indicator, and calculate the objective weight corresponding to the indicator based on the correlation coefficient; S303, after receiving the subjective weight and objective weight of each indicator, construct multiple conditional functions based on these data, conduct joint analysis on the conditional functions to obtain weight coefficients, and calculate the corresponding combined weight according to the weight coefficients and the subjective weight and objective weight of each indicator; S304, dividing the risk level into four levels, from level 1 to level 4, namely, negligible risk, acceptable risk, acceptable risk after repair, and unacceptable risk, and setting a corresponding characteristic coefficient for each risk level; S305. After analyzing multiple risk parameters corresponding to each indicator based on the basic indicator data collected in the designated mining area, determine the characteristic coefficient corresponding to each risk parameter, combine it with the combined weight of the indicator, and thus obtain the characteristic coefficient of each secondary indicator. Follow the same steps to determine the characteristic coefficient of the primary indicator, calculate the similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level, take the risk level with the highest similarity as the actual risk level of the indicator, and after determining the risk levels of all indicators, output them through a visual interface.

6. A coal mine safety monitoring data fusion processing method according to claim 1, characterized in that: The step S4 specifically comprises the following steps: S401, using a line scanning laser to scan the coal mine tunnel, and using a camera to photograph the light stripes generated by the scanning, and transmitting the obtained multiple image data to a database; S402, after extracting the light streak image data, distinguish all the pixels in the image according to the set color component range, store the pixels that meet the set range into a target set, and after determining the area formed by all the pixels in the set, use the area as the target area; S403, filter the target area using a 3×3 size template, perform grayscale binarization on the processed image data, determine the foreground point and the background point according to the binarized image data, and construct an inscribed circle in the non-background point area with each foreground point as the center point. If the current inscribed circle belongs to a subset of other inscribed circles, no operation is performed, otherwise the corresponding center point is stored in the key point set; S404, after traversing each pixel in the set, calculate the distance value between the pixel and the nearest background point, and use the value to replace the original pixel value; S405, after setting the size of the local neighborhood, compare the value of each pixel in the set with the values ​​of other foreground points in the area, and if the distance value of the current pixel in any direction is the largest, determine the pixel as the point to be grown; S406, after counting all the points to be grown in the target area, determine the number of foreground points in the neighborhood around the point to be grown, if the number is lower than the quantity threshold, set the pixel value corresponding to the point to be grown to zero, otherwise do not perform any operation.

7. A coal mine safety monitoring data fusion processing method according to claim 1, characterized in that: The step S5 specifically comprises the following steps: S501, analyzing the refined target area according to the position of the point to be grown, traversing each pixel point in the area row by row, determining the position of the pixel point whose gray value is greater than the preset value in each row, calculating the position of the center point corresponding to the row according to the gray value and position of these pixel points, and counting all the center points and specific positions in the area and storing them in the database; S502, after obtaining the focal length and scale factor of the camera, using a coordinate conversion algorithm to calculate the actual coordinates according to the pixel coordinates corresponding to each center point; S503, after determining the center point and specific position of all image data, use 3D software to splice these data to obtain a complete point cloud map, and perform surface fitting on it to determine the simulated 3D surface corresponding to the coal mine tunnel; S504, extract the initial three-dimensional surface of the coal mine roadway, slice the initial three-dimensional surface and the simulated three-dimensional surface to obtain multiple initial point cloud slices and simulated point cloud slices, compare the initial point cloud slices and the simulated point cloud slices according to the slice position, analyze the specific position and volume of the roadway deformation point, use deep learning technology to analyze these data, and determine the deformation level of the current roadway.

8. A coal mine safety monitoring data fusion processing system, characterized in that: The data fusion processing system is applicable to a coal mine safety monitoring data fusion processing method according to any one of claims 1 to 7, comprising an indicator generation unit (1), an effective indicator determination unit (2), a risk level output unit (3), a to-be-grown point screening unit (4), a deformation point analysis unit (5) and an early warning prompt unit (6); The indicator generation unit (1) constructs a plurality of basic indicators using a plurality of roof accident analysis cases, performs cluster analysis on all the basic indicators, determines the secondary indicators and primary indicators corresponding to each basic indicator, uses ten experts to score each indicator according to four dimensions, and marks invalid indicators and valid indicators according to the scores; The effective index determination unit (2) obtains data corresponding to each index in multiple mining areas, determines the number of important indexes based on these data, uses a component evaluation algorithm to analyze the component values ​​corresponding to the overall index and the effective index, uses these values ​​to further determine the number of effective indexes, and deletes invalid indicators and corresponding data; After the risk level output unit (3) determines the weight coefficient, the subjective weight and the objective weight, it calculates the corresponding combined weight according to these data, divides the risk level into four levels, and sets a corresponding characteristic coefficient for each risk level. According to the basic indicator data collected in the designated mining area and the combined weight of the indicator, the characteristic coefficient corresponding to each indicator is analyzed, and the similarity between the characteristic coefficient of each indicator and the corresponding characteristic coefficient in the risk level is calculated. After the risk level of all indicators is determined according to the similarity, it is output through a visual interface; The to-be-grown point screening unit (4) extracts light streak image data of the coal mine tunnel, determines the corresponding target area in the image data, calculates the distance value between each pixel point in the area and the nearest background point, uses the value to replace the original pixel value, and screens out the to-be-grown points based on these data; The deformation point analysis unit (5) analyzes the refined target area according to the position of the point to be grown, counts all the center points and specific positions in the area, and then uses three-dimensional software to splice these data to obtain a simulated three-dimensional surface corresponding to the coal mine tunnel, and analyzes the specific position and volume of the tunnel deformation occurrence point; If the coal mine roof monitoring shows an indicator with a risk level of four and the tunnel deformation level is a severe deformation, the early warning prompt unit (6) will directly report to the control center and send out a prompt message by voice broadcasting; if the number of first-level indicators with a risk level of three in the coal mine roof monitoring is greater than one, it will directly report to the control center; if the number of first-level indicators with a risk level of three in the coal mine roof monitoring is less than or equal to one and the tunnel deformation level is a slight deformation, a prompt will be sent out by voice broadcasting.

Images