A coal yard safety analysis system based on infrared imaging and point cloud data fusion

Abstract

The present invention discloses a coal yard safety analysis system based on the fusion of infrared imaging and point cloud data, which specifically relates to the field of safety analysis technology. The system collects infrared imaging data and three-dimensional point cloud data on the surface of a coal pile, constructs an initial emissivity distribution map, divides the map into multiple local grid areas, and extracts image texture features and surface structure information of each local grid area. A corrected emissivity distribution map is generated by local dynamic compensation of the initial emissivity. Abnormal areas are identified based on the corrected emissivity map and marked as potential risk areas. By constructing a heat transfer diffusion model, the time dependence of heat diffusion is evaluated. The fractal analysis method is used to evaluate the degree of balance of heat distribution in potential risk areas. Based on the time dependence and spatial heterogeneity of heat diffusion, the safety risk level of the coal yard is evaluated, and coal yard safety warning information is generated, thereby realizing emissivity correction and accurate thermal risk assessment of multi-state coal in mixed coal piles.

Description

Technical Field

[0001] The present invention relates to the technical field of safety analysis, and more specifically, to a coal yard safety analysis system based on the fusion of infrared imaging and point cloud data. Background Art

[0002] Infrared thermal imaging cameras are often used in coal yard safety monitoring to monitor the surface temperature of coal piles to identify potential spontaneous combustion hazards or areas of thermal anomalies. The accuracy of infrared temperature measurement is highly dependent on the surface emissivity of the object being measured. Coal, a material with complex composition and variable state, has a significantly different emissivity in real-world scenarios depending on type, batch, degree of weathering, and moisture content, exhibiting spatial inhomogeneity and dynamic temporal variations.

[0003] In the existing technology, a unified emissivity setting method is usually adopted without combining point cloud data. When processing mixed coal piles, it will lead to significant deviations in local temperature estimation and even affect the accuracy of judging thermal anomaly risks.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a coal yard safety intelligent analysis method and system based on the fusion of infrared imaging and point cloud data to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A coal yard safety analysis system based on infrared imaging and point cloud data fusion, including:

[0008] The data acquisition module collects infrared imaging data and three-dimensional point cloud data on the surface of the coal pile, performs multi-source information fusion processing on the infrared imaging data and the three-dimensional point cloud data, and generates an initial emissivity distribution map of the coal pile surface;

[0009] The meshing module divides the coal pile surface into multiple local mesh areas based on the initial emissivity distribution map, and extracts the texture features and surface structure information of each local mesh area;

[0010] The emission correction module performs local dynamic compensation on the initial emissivity distribution map based on texture features and surface structure information to generate a corrected emissivity distribution map;

[0011] The anomaly detection module identifies abnormal areas on the coal pile surface based on the corrected emissivity distribution map and marks them as potential risk areas;

[0012] The diffusion assessment module analyzes the heat distribution characteristics in the potential risk area through the heat transfer diffusion model and evaluates the time dependence of heat diffusion;

[0013] The fractal assessment module uses fractal analysis to assess the degree of heat distribution balance in potential risk areas;

[0014] The risk warning module evaluates the safety risk level of the coal yard based on the time dependence of heat diffusion and the balance of heat distribution, and generates coal yard safety warning information.

[0015] In a preferred embodiment, infrared imaging data and three-dimensional point cloud data are collected on the surface of the coal pile, and multi-source information fusion processing is performed on the infrared imaging data and the three-dimensional point cloud data to generate an initial emissivity distribution map of the coal pile surface, specifically:

[0016] Infrared thermal imaging equipment is used to collect infrared imaging data of the coal pile surface to obtain an infrared radiation intensity distribution image;

[0017] Use 3D laser scanning equipment to collect spatial position coordinate data and spatial surface structure feature data of the coal pile surface to obtain 3D point cloud data of the coal pile surface;

[0018] Perform spatial coordinate alignment and time stamp alignment on the infrared radiation intensity distribution image and the three-dimensional point cloud data, and establish a mapping relationship between the infrared image pixel points and the point cloud coordinate points;

[0019] Based on the mapping relationship between infrared image pixels and point cloud coordinates, the initial emissivity distribution map of the coal pile surface is generated.

[0020] In a preferred embodiment, based on the initial emissivity distribution map, the coal pile surface is divided into multiple local grid areas, and the texture features and surface structure information of each local grid area are extracted, specifically:

[0021] The local spatial boundary of the coal pile surface is determined based on the initial emissivity distribution map, and the coal pile surface is divided into multiple local grid areas;

[0022] For each local grid area, surface structure information is extracted based on 3D point cloud data;

[0023] For each local grid area, texture features are extracted based on the infrared radiation intensity distribution image.

[0024] In a preferred embodiment, local dynamic compensation is performed on the initial emissivity distribution map based on texture features and surface structure information to generate a modified emissivity distribution map, specifically:

[0025] Establishing a mapping relationship between each local grid area in the initial emissivity distribution map and the corresponding texture features and surface structure information;

[0026] In each local grid area, a spatial structure model is constructed based on the three-dimensional point cloud data;

[0027] In each local grid area, an image texture feature model is constructed based on the infrared radiation intensity distribution image;

[0028] The spatial structure model is jointly fitted with the image texture feature model to obtain the emissivity adjustment factor of each local grid area;

[0029] Based on the emissivity adjustment factor of each local grid area, the original emissivity value is locally dynamically compensated to generate a corrected emissivity distribution map.

[0030] In a preferred embodiment, abnormal areas on the surface of the coal pile are identified based on the modified emissivity distribution map and marked as potential risk areas, specifically:

[0031] Perform spatial neighborhood analysis on the corrected emissivity distribution map to determine the local statistical threshold of emissivity in each local grid area;

[0032] Determine the abnormal deviation degree of the emissivity in each local grid area based on the local statistical threshold, and generate emissivity abnormal intensity data;

[0033] Determine the spatial position boundary of the abnormal area on the surface of the coal pile based on the abnormal emissivity intensity data;

[0034] The abnormal area on the coal pile surface is divided into spatial connectivity to determine the overall spatial range of the abnormal area;

[0035] The overall spatial range of the abnormal area is marked to obtain the potential risk area.

[0036] In a preferred embodiment, the heat transfer diffusion model is used to analyze the heat distribution characteristics in the potential risk area and evaluate the time dependence of heat diffusion, specifically:

[0037] Establish a heat transfer and diffusion model based on the overall spatial scope of the potential risk area;

[0038] Conduct numerical simulation of heat diffusion in the potential risk area based on the heat transfer diffusion model to determine the heat distribution data of each local grid area in the potential risk area;

[0039] Analyze the heat diffusion trend based on the heat distribution data of each local grid area in the potential risk area;

[0040] According to the heat diffusion trend, the time dependence of heat diffusion in the potential risk area is evaluated to obtain the time characteristic data of heat diffusion.

[0041] In a preferred embodiment, the degree of heat distribution balance in the potential risk area is evaluated by a fractal analysis method, specifically:

[0042] Extract heat distribution data of each local grid area within the potential risk area;

[0043] Mapping the heat distribution data of each local grid area within the potential risk area into a spatial heat distribution matrix;

[0044] Multi-scale fractal dimension calculation is performed on the spatial heat distribution matrix to obtain the fractal dimension sequence of potential risk areas at different spatial scales;

[0045] According to the changing trend of the fractal dimension sequence, the spatial balance of heat distribution in the potential risk area is evaluated, and the spatial characteristic data of heat diffusion are obtained.

[0046] In a preferred embodiment, based on the time dependency of heat diffusion and the degree of balance of heat distribution, the safety risk level of the coal yard is assessed and coal yard safety warning information is generated, specifically:

[0047] Based on the heat diffusion time characteristic data and heat diffusion space characteristic data, the feature vector is spliced to obtain the comprehensive risk assessment input data set.

[0048] Based on the preset coal yard risk assessment model and combined with the comprehensive risk assessment input data set, the risk level assessment of the potential risk area is performed and the risk level value of the potential risk area is output;

[0049] Classify the potential risk areas into risk levels according to the risk level values and generate a risk level map of the potential risk areas;

[0050] Generate corresponding coal yard safety warning information based on the risk level map.

[0051] The technical effects and advantages of the coal yard safety analysis system based on infrared imaging and point cloud data fusion of the present invention are as follows:

[0052] By fusing infrared imaging data with three-dimensional point cloud data, an initial emissivity distribution map is generated, which improves the basic characterization accuracy of the thermal characteristics of the coal pile. By dividing the coal pile into local grid areas and extracting texture and structural features, the perception of fine-grained spatial differences is achieved. The emissivity is dynamically compensated based on local characteristics, which solves the temperature measurement deviation problem caused by a single emissivity setting. The accuracy of high-temperature hazard positioning is enhanced by identifying abnormal areas through the corrected emissivity map. The time dependency is analyzed using a heat transfer diffusion model to predict the development trend of risks. Fractal analysis is used to evaluate spatial heterogeneity, which improves the ability to distinguish complex thermal distribution structures. Finally, the risk level of the coal yard is assessed based on the time dependency of heat diffusion and the degree of balance of heat distribution, and early warning information is generated, enhancing the response capability and early recognition capability to abnormal conditions of complex coal piles. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a structural schematic diagram of a coal yard safety analysis system based on infrared imaging and point cloud data fusion in the present invention. DETAILED DESCRIPTION

[0054] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0055] Example 1

[0056] Figure 1 The present invention provides a coal yard safety analysis system based on infrared imaging and point cloud data fusion, comprising:

[0057] The data acquisition module collects infrared imaging data and three-dimensional point cloud data on the surface of the coal pile, performs multi-source information fusion processing on the infrared imaging data and the three-dimensional point cloud data, and generates an initial emissivity distribution map of the coal pile surface;

[0058] The meshing module divides the coal pile surface into multiple local mesh areas based on the initial emissivity distribution map, and extracts the texture features and surface structure information of each local mesh area;

[0059] The emission correction module performs local dynamic compensation on the initial emissivity distribution map based on texture features and surface structure information to generate a corrected emissivity distribution map;

[0060] The anomaly detection module identifies abnormal areas on the coal pile surface based on the corrected emissivity distribution map and marks them as potential risk areas;

[0061] The diffusion assessment module analyzes the heat distribution characteristics in the potential risk area through the heat transfer diffusion model and evaluates the time dependence of heat diffusion;

[0062] The fractal assessment module uses fractal analysis to assess the degree of heat distribution balance in potential risk areas;

[0063] The risk warning module evaluates the safety risk level of the coal yard based on the time dependence of heat diffusion and the balance of heat distribution, and generates coal yard safety warning information.

[0064] Specifically, infrared imaging data and three-dimensional point cloud data are collected on the coal pile surface, and multi-source information fusion processing is performed on the infrared imaging data and the three-dimensional point cloud data to generate an initial emissivity distribution map of the coal pile surface, including:

[0065] Infrared thermal imaging equipment is used to collect infrared imaging data of the coal pile surface to obtain an infrared radiation intensity distribution image;

[0066] Specifically, a high-resolution infrared thermal imaging device is used to continuously capture the coal pile surface at a fixed sampling frequency, producing an infrared radiation intensity image covering all areas of the coal pile surface. At least five image samples are taken at each location, and the average of these five image samples is calculated to obtain a numerical infrared radiation intensity distribution value for each location. This reduces random errors caused by environmental factors and improves data reliability. The data captured by the infrared thermal imaging device is a two-dimensional infrared radiation intensity distribution image. Each pixel in the infrared radiation intensity distribution image corresponds to a specific location on the coal pile surface in real space.

[0067] Use 3D laser scanning equipment to collect spatial position coordinate data and spatial surface structure feature data of the coal pile surface to obtain 3D point cloud data of the coal pile surface;

[0068] Specifically, a 3D laser scanner emits laser pulses at the coal pile surface, measures the time of flight from emission to return, and then calculates the actual spatial distance to the corresponding location on the coal pile surface by multiplying the speed of light by the time of flight and dividing it by two. The laser scanner scans with millimeter-level spatial accuracy, generating 3D coordinate data for each scanned location. Continuously scanning all areas of the coal pile surface generates a point cloud of 3D coordinate data.

[0069] Perform spatial coordinate alignment and time stamp alignment on the infrared radiation intensity distribution image and the three-dimensional point cloud data, and establish a mapping relationship between the infrared image pixel points and the point cloud coordinate points;

[0070] Specifically, the infrared radiation intensity distribution image is two-dimensional spatial data, and the point cloud data is three-dimensional spatial data. Therefore, a spatial mapping algorithm is needed to correspond the pixels of the two-dimensional infrared radiation intensity distribution image to the three-dimensional spatial coordinates. The spatial mapping algorithm is specifically as follows: according to the spatial coordinates of the three-dimensional point cloud data and the scanning angle parameters preset by the three-dimensional laser scanning device, each spatial coordinate point is projected onto the two-dimensional image plane to determine the correspondence between each point cloud coordinate point and the infrared image pixel position. A high-precision time synchronization protocol is used to ensure that the acquisition of infrared images and point cloud data is completed at the same time node, and each spatial coordinate point has the same timestamp as the infrared image pixel point to ensure the effectiveness of spatial alignment and temporal synchronization. Through the dual calibration of spatial coordinates and timestamps, the data matching accuracy in the fusion process can be ensured.

[0071] Based on the mapping relationship between infrared image pixels and point cloud coordinates, the initial emissivity distribution map of the coal pile surface is generated;

[0072] Specifically, the calculation of the initial emissivity is based on the theory of infrared radiation of objects. The specific emissivity value is determined by the relationship between infrared radiation intensity, the surface temperature of the coal pile, and the ambient temperature. Specifically, the infrared radiation intensity distribution image is used to obtain the infrared radiation intensity of each pixel. Based on the spatial characteristics of the coal pile surface determined by the point cloud number and the temperature distribution at the corresponding position, the temperature distribution of the coal pile surface and the ambient temperature conditions are considered. The principle of conservation of infrared radiation energy is used for conversion to obtain the initial emissivity value for each spatial position. The specific calculation method is as follows: the initial emissivity value is equal to the ratio of the infrared radiation intensity to the blackbody radiation intensity value at the corresponding surface position. The blackbody radiation intensity value is calculated by combining the temperature value of the corresponding surface position with Planck's radiation law. The initial emissivity value of each position on the coal pile surface is obtained through the above method, and a complete initial emissivity distribution map of the coal pile surface is drawn according to the spatial coordinate position.

[0073] Specifically, based on the initial emissivity distribution map, the coal pile surface is divided into multiple local grid areas, and the texture features and surface structure information of each local grid area are extracted, including:

[0074] The local spatial boundary of the coal pile surface is determined based on the initial emissivity distribution map, and the coal pile surface is divided into multiple local grid areas;

[0075] Specifically, the spatial coordinate range corresponding to the coal pile surface is obtained based on the initial emissivity distribution map, and the minimum bounding rectangle is determined. The minimum bounding rectangle is divided into two-dimensional grids according to the equal-interval division rule to ensure that the coal pile surface covered by each grid area is continuous in spatial distribution and has no overlap. During the division process, the minimum side length of a single grid area is set to several times the minimum effective ranging unit of the three-dimensional laser scanning device to avoid the loss of structural details due to insufficient spatial resolution. The grid numbers after division are identified using a row and column numbering system, and the initial emissivity value set corresponding to each grid area is extracted as input data for feature analysis.

[0076] For each local grid area, surface structure information is extracted based on 3D point cloud data;

[0077] Specifically, in each divided local grid area, all three-dimensional point cloud coordinate data within the coverage area of the local grid area are extracted. Local geometric structure calculations are performed within each set of three-dimensional point cloud coordinate data to obtain the spatial structural feature information of the surface of the local grid area:

[0078] Calculate the direction of the spatial normal vector: Within each point cloud subset, the least squares plane fitting method is used to perform local plane approximation modeling on the point set, and then the regional normal vector is determined based on the vertical direction vector of the fitting plane to reflect the inclination direction of the coal pile surface in this area.

[0079] Calculate the local curvature change value: Calculate the surface curvature of the grid area based on the change amplitude of the spatial normal vector distribution. The curvature calculation adopts the weighted average method based on the change amount of the normal vector angle of the point cloud tangent plane. The more drastic the angle change, the higher the curvature, reflecting whether there are abnormal protrusions or depressions on the surface structure of the coal pile.

[0080] Statistical point cloud density: Count the number of point clouds per unit area to determine the structural information integrity and visibility of the area. Areas with too low point cloud density will be marked as incomplete features.

[0081] For each local grid area, texture features are extracted based on the infrared radiation intensity distribution image;

[0082] Specifically, grayscale normalization is performed on the image area corresponding to each local grid area in the initial emissivity distribution map to eliminate errors caused by device response offset and brightness non-uniformity.

[0083] Based on the standard texture analysis method, the following feature extraction operations are performed on each local grid area in the image:

[0084] Calculate the gray level co-occurrence matrix: construct a two-dimensional gray level relationship matrix based on the gray level difference of the image to reflect the gray level correlation between pixels in the area;

[0085] Extract texture contrast parameters: Calculate the image clarity change trend based on the weighted average of the grayscale differences in the grayscale co-occurrence matrix;

[0086] Extract entropy parameter: measures the randomness of the grayscale distribution within the region. The higher the entropy value, the more complex the internal structure of the region.

[0087] Calculate the homogeneity parameter: reflects the degree of grayscale similarity in the image and is used to identify whether there is local radiation anomaly in the area.

[0088] Specifically, local dynamic compensation is performed on the initial emissivity distribution map based on texture features and surface structure information to generate a corrected emissivity distribution map, including:

[0089] Establishing a mapping relationship between each local grid area in the initial emissivity distribution map and the corresponding texture features and surface structure information;

[0090] Specifically, based on the local grid division, for each clearly numbered local grid area, the corresponding image texture features and spatial structure features are combined and encoded to form a feature map. The feature map uses the local grid number as an index and includes: the set of infrared radiation image texture feature parameters (including texture contrast parameter, entropy parameter, homogeneity parameter) and the set of spatial point cloud structure parameters (including spatial normal vector direction, local curvature change value, point cloud density) for the grid area, as well as the original emissivity value of the area in the initial emissivity distribution map.

[0091] In each local grid area, a spatial structure model is constructed based on the three-dimensional point cloud data;

[0092] Specifically, in each grid area, the spatial structure model is constructed using the spatial normal vector direction and local curvature as the main structural parameters. The specific process is as follows: a directional tensor matrix is constructed based on the point cloud normal vector, and the tensor matrix is subjected to eigenvalue decomposition to extract the distribution of main direction changes to reflect the geometric trend of the coal pile surface in the grid area. At the same time, combined with the curvature change, the direction tensor and curvature index are weighted and combined according to a specific weight rule to form a spatial structure model of the grid area. The model can be used to quantify whether the coal pile surface in this area has complex landform features, such as protrusions, depressions, or folds.

[0093] In each local grid area, an image texture feature model is constructed based on the infrared radiation intensity distribution image;

[0094] Specifically, an image texture model is constructed based on texture feature parameters. A multidimensional feature space for local image texture is constructed based on texture contrast, entropy, and homogeneity parameters extracted from the gray-level co-occurrence matrix. Principal component analysis is used to reduce the dimensionality of these parameters, forming a set of primary texture expression factors. These primary texture expression factors are used to characterize the typical variation patterns of the infrared image of that region in visual space.

[0095] The spatial structure model is jointly fitted with the image texture feature model to obtain the emissivity adjustment factor of each local grid area;

[0096] Specifically, a weighted regression model was used to establish a joint mapping regression equation, using the output values of the spatial structure model and the main texture expression factors of the image texture feature model as independent variables, and the emissivity value of the region in the initial emissivity distribution map as the dependent variable. A minimum residual optimization process was used to determine the adjustment weight of the characteristic response of the region to the emissivity deviation value. Finally, the regression equation was used to calculate the emissivity adjustment factor for the grid region. The magnitude of the emissivity adjustment factor represents the actual responsiveness of the current grid structure and image features in the current thermal field and reflects the degree of correction to the original emissivity value.

[0097] Based on the emissivity adjustment factor of each local grid area, the original emissivity value is locally dynamically compensated to generate a corrected emissivity distribution map;

[0098] Specifically, for each local grid area, the corrected emissivity is calculated by multiplying the original emissivity value by the emissivity adjustment factor for that area to obtain the initial compensation result for that area. A spatial smoothing factor is introduced to balance the emissivity transition difference between that area and its directly adjacent local grid areas. The spatial smoothing factor is equal to the difference between the average of the original emissivity of the eight directly adjacent local grid areas and the original emissivity of the area. The corrected emissivity value is the weighted average of the initial compensation result and the spatial smoothing factor.

[0099] While achieving local compensation, the continuity of the emissivity transition between regions is considered to avoid single-point mutations. Finally, the emissivity values of all compensated local grid areas are combined to draw a complete corrected emissivity distribution map of the coal pile surface.

[0100] Specifically, abnormal areas on the coal pile surface are identified based on the corrected emissivity distribution map and marked as potential risk areas, including:

[0101] Perform spatial neighborhood analysis on the corrected emissivity distribution map to determine the local statistical threshold of emissivity in each local grid area;

[0102] Specifically, the corrected emissivity distribution map of the coal pile surface was divided into multiple local grid regions. Each region consists of a central grid and its eight adjacent grids, for a total of nine spatial units. For each central grid region, the emissivity values of the central grid region and its eight adjacent regions were extracted. The statistical mean and standard deviation of the nine emissivity values were calculated. The sum of the statistical mean and standard deviation was used as the local statistical threshold for the central grid region.

[0103] Determine the abnormal deviation degree of the emissivity in each local grid area based on the local statistical threshold, and generate emissivity abnormal intensity data;

[0104] Specifically, for each local grid area, the difference between the corrected emissivity value and the local statistical threshold is calculated. If the difference is positive, it indicates that the emissivity of the area has exceeded the normal fluctuation range, and the difference is the emissivity anomaly intensity. If the difference is zero or negative, it indicates that the area has not exceeded the local threshold, and the emissivity anomaly intensity is set to zero.

[0105] The anomaly intensity values corresponding to all local grid regions are spatially spliced to generate an emissivity anomaly intensity data map covering the entire coal pile surface. This emissivity anomaly intensity data map serves as a spatial basis for quantifying the degree of thermal anomaly and provides a foundation for locating the anomaly boundary.

[0106] Determine the spatial position boundary of the abnormal area on the surface of the coal pile based on the abnormal emissivity intensity data;

[0107] Specifically, we traverse all grids in the emissivity anomaly intensity data map and identify a set of regions with anomaly intensities greater than zero. These regions are marked as initial anomaly region candidates. Adjacent anomaly candidate points are recursively merged into spatially connected regions using an eight-neighborhood algorithm to form a preliminary set of anomaly blocks.

[0108] In the abnormal block set, the boundary position coordinates of each block are calculated, including the minimum horizontal number, maximum horizontal number, minimum vertical number, and maximum vertical number, to form a two-dimensional bounding box of the abnormal area on the emissivity image, which serves as the actual position range of the abnormal area in the image space.

[0109] The abnormal area on the coal pile surface is divided into spatial connectivity to determine the overall spatial range of the abnormal area;

[0110] Specifically, based on the initial set of bounding boxes, the bounding boxes are analyzed for overlap. If there is spatial overlap or boundary contact between two bounding boxes, they are merged into a new anomaly region. The merging process uses the extended boundary comparison method, which determines whether two bounding boxes have a consecutive numbered intersection in the horizontal or vertical direction. If so, the two bounding boxes are logically merged into a single spatial entity.

[0111] After completing the connectivity judgment of all abnormal block sets, the final abnormal area set is output. Each abnormal area has a clear number identification and spatial boundary range, including its internal emissivity abnormal intensity data, including maximum value, mean value and area.

[0112] Mark the overall spatial range of the abnormal area to obtain the potential risk area;

[0113] Specifically, each subset of abnormal regions determined through mutual overlap analysis is marked as a potential risk region.

[0114] Specifically, the heat transfer diffusion model is used to analyze the heat distribution characteristics in the potential risk area and evaluate the time dependence of heat diffusion, including:

[0115] Establish a heat transfer and diffusion model based on the overall spatial scope of the potential risk area;

[0116] Specifically, after the potential risk areas are identified, the two-dimensional boundary range corresponding to each potential risk area is extracted. This boundary range is used as the calculation domain boundary to define the spatial boundary conditions of the model calculation potential risk areas. The definition principle of spatial boundary is:

[0117] The spatial boundary should cover the entire potential risk area and extend one unit grid outward to introduce the boundary heat diffusion effect;

[0118] The emissivity value and temperature distribution in the boundary area are jointly determined based on the corrected emissivity map and the externally collected temperature data to ensure the accuracy of the initial boundary conditions of the model.

[0119] A two-dimensional steady-state heat conduction model was established within the defined spatial boundaries. The model assumes that heat diffusion on the coal pile surface is primarily in-plane, and that vertical heat exchange can be simplified to a steady background heat flux, allowing for approximate modeling on a spatial plane. The heat diffusion process was discretized using the finite difference method. The heat conduction differential equations were approximated at each grid node to form a system of linear equations that describe the temporal evolution of heat transfer.

[0120] Conduct numerical simulation of heat diffusion in the potential risk area based on the heat transfer diffusion model to determine the heat distribution data of each local grid area in the potential risk area;

[0121] Specifically, after the heat conduction model is discretized, a heat diffusion simulation is performed on the model. The initial conditions of the simulation are set based on the corrected emissivity map and the infrared temperature distribution map. The initial heat value of each local grid is calculated by combining the corrected emissivity value and the actual temperature value of the area:

[0122] The initial heat value is equal to the emissivity value multiplied by the surface temperature, multiplied by the unit specific heat and the unit thickness.

[0123] The simulation process uses an explicit time iteration method with equally spaced time steps. At each time step, the temperature of each grid cell is updated using the difference scheme of the heat conduction equation, which in turn updates the heat value of each cell. Each time step update completes, a snapshot of the heat distribution of the entire grid in the potential risk area is obtained.

[0124] The entire simulation cycle lasts for several time steps until the heat diffusion process reaches a preset equilibrium state, meaning that the temperature change rate of all grid cells falls below a specified threshold. After the simulation ends, the complete heat distribution data sequence corresponding to each time step is output to form a heat time evolution matrix.

[0125] Analyze the heat diffusion trend based on the heat distribution data of each local grid area in the potential risk area;

[0126] Specifically, using the heat distribution data at each time step in the heat time evolution matrix, a time series difference calculation is performed on each local grid to obtain a sequence of heat change values per unit time. Based on this change value sequence, the trend of the heat diffusion rate of each grid is statistically analyzed.

[0127] The specific method is: using the square sum of the heat difference as the calculation indicator, calculate the total heat change amplitude of each grid; then calculate the change direction of the total heat change amplitude over time during the entire simulation cycle to determine whether the heat diffusion is accelerating, slowing down or tending to a stable state.

[0128] Based on the heat diffusion trend, the time dependence of heat diffusion in the potential risk area is evaluated to obtain the time characteristic data of heat diffusion;

[0129] Specifically, the time dependency assessment is based on trend analysis data to comprehensively judge the degree of correlation between heat changes and time. The assessment method is:

[0130] For each local grid cell, the heat value sequence at multiple consecutive time steps is extracted. The linear correlation coefficient between the heat value and the time step number is calculated using the Pearson correlation coefficient algorithm. The correlation coefficient value is between zero and one, indicating the monotonicity of the heat diffusion process in the grid cell. A correlation coefficient close to one indicates strong monotonic diffusion, while a correlation coefficient close to zero indicates large fluctuations in the diffusion direction.

[0131] The sliding extreme value identification algorithm is applied to each local grid heat time series. A fixed time window width is set to identify the local maximum and local minimum positions in the sequence, and each trend reversal is defined as an inflection point. The total number of inflection points in the entire heat diffusion process is counted to reflect whether there are drastic fluctuations or stage-by-stage mutations in the heat diffusion process.

[0132] The comprehensive linear correlation coefficient and the total number of inflection points are used as the time characteristic data of the time-dependent heat diffusion in the potential risk area.

[0133] Specifically, the fractal analysis method is used to evaluate the degree of heat distribution balance in the potential risk area, including:

[0134] Extract heat distribution data of each local grid area within the potential risk area;

[0135] Specifically, from the heat time evolution matrix, the steady-state time step after the heat diffusion process ends is selected to extract the heat values for all local grids within the potential risk area at that moment. Each local grid area corresponds to a unique heat value, representing the level of heat energy accumulation in the final thermal equilibrium stage.

[0136] Mapping the heat distribution data of each local grid area within the potential risk area into a spatial heat distribution matrix;

[0137] Specifically, the value of each element in the spatial heat distribution matrix represents the calorific value of the corresponding grid, and each row and column corresponds to the spatial coordinates of the coal pile surface. The spatial heat distribution matrix is normalized to normalize all calorific values to between zero and one.

[0138] Apply a heat grading process to the standardized spatial heat distribution matrix. Divide the entire value range into several graded segments, for example, categorizing heat values into low, medium, and high heat zones. Each grade corresponds to a grayscale code, forming a heat distribution grade image.

[0139] Multi-scale fractal dimension calculation is performed on the spatial heat distribution matrix to obtain the fractal dimension sequence of potential risk areas at different spatial scales;

[0140] Specifically, the box dimension method is used to perform multi-scale box coverage analysis on the heat distribution level image:

[0141] Set multiple square box grids of different sizes on the heat distribution level image, starting from the smallest cell, and expand the box size by a fixed ratio each time to form a series of multi-scale box partition sets;

[0142] At each scale, the heat level image is covered with a box grid, and the number of boxes in the current box grid that contain non-empty pixels (i.e., there are high heat level points) is counted; high heat level points are the representation units of high heat areas in the heat distribution level image, and they correspond to each other.

[0143] Record the box size and the corresponding number of non-empty boxes at each scale, and plot the double-logarithmic graph of the number of boxes and box size in a logarithmic coordinate system for linear regression analysis;

[0144] The absolute value of the slope of the regression line is the fractal dimension value of the current heat level image, which indicates the complexity of the spatial heat structure at the current scale.

[0145] Repeat the above process to calculate the fractal dimensions at multiple scales and obtain the fractal dimension sequence of the potential risk area.

[0146] According to the changing trend of the fractal dimension sequence, the spatial balance of heat distribution in the potential risk area is evaluated to obtain the spatial characteristic data of heat diffusion;

[0147] Specifically, the spatial feature data include: fractal mean index: the sum of the fractal dimension values corresponding to all scales in the fractal dimension sequence is divided by the sequence length to obtain the average fractal dimension of the potential risk area, which is used to quantify the complexity of the overall spatial distribution;

[0148] Fractal Fluctuation Index: Calculates the difference between the maximum and minimum fractal dimension values and uses it as the fluctuation amplitude to quantify the intensity of changes in fractal complexity at different spatial scales.

[0149] Main scale position: In the fractal dimension sequence, the scale number corresponding to the maximum fractal dimension value is taken as the main scale position, indicating that the heat distribution is the most uneven at this spatial scale;

[0150] Balance index: The balance index is calculated by dividing the fractal mean index by the fractal fluctuation index to reflect the stability of the spatial distribution. The larger the balance index, the smoother the change in the overall fractal complexity and the more balanced the spatial structure.

[0151] Specifically, based on the time dependence of heat diffusion and the degree of heat distribution balance, the safety risk level of the coal yard is assessed and coal yard safety warning information is generated, including:

[0152] Based on the heat diffusion time characteristic data and the heat diffusion space characteristic data, feature vector splicing is performed to obtain the comprehensive risk assessment input data set;

[0153] Specifically, the heat diffusion time characteristic data and the heat diffusion space characteristic data are combined into a feature vector, and a feature vector sample of each potential risk area is constructed to obtain a comprehensive risk assessment input data set.

[0154] Based on the preset coal yard risk assessment model and combined with the comprehensive risk assessment input data set, the risk level assessment of the potential risk area is performed and the risk level value of the potential risk area is output;

[0155] Specifically, the coal yard risk assessment model adopts a multi-dimensional input risk mapping structure, which is as follows: the input feature vector is mapped to a set of risk score intervals, and the potential risk level value is output according to the linear weighted assessment rule.

[0156] Each feature in the risk scoring model has a preset weight parameter. These weights are determined based on training results from historically annotated samples. During the training phase, a minimum mean square error optimization strategy is used to adjust these parameters, ensuring the output risk level has the smallest deviation from the manually annotated level. The training sample data is derived from historical examples of typical coal pile safety monitoring and undergoes standard feature normalization preprocessing.

[0157] The model calculation method is: the heat diffusion time feature subvector and the heat diffusion space feature subvector are jointly input, multiplied by the corresponding risk weight coefficient respectively, and then summed up. The value is normalized to a risk level value between zero and one through a first-order activation function.

[0158] The output result is the risk level value of each potential risk area, which is used for risk level classification and risk level map construction.

[0159] Classify the potential risk areas into risk levels according to the risk level values and generate a risk level map of the potential risk areas;

[0160] Specifically, the risk level values of all potential risk areas are divided into three levels:

[0161] If the risk level value is greater than or equal to 70%, it is classified as a high-level potential risk area;

[0162] If the risk level value is greater than or equal to 40% and less than 70%, it is classified as a medium-level potential risk area;

[0163] If the risk level value is less than 40%, it is classified as a low-level potential risk area.

[0164] Each potential risk area is assigned a corresponding risk level mark according to the above classification standards and superimposed on the two-dimensional spatial map of the coal pile surface to form a spatial distribution map of the potential risk area, which is the risk level map.

[0165] Generate corresponding coal yard safety warning information based on the risk level map;

[0166] Specifically, based on the risk level map, security warning information is generated for different levels of potential risk areas, including:

[0167] Risk level classification label (high, medium, low);

[0168] Recommend response measures (e.g., local retesting, activating cooling system).

[0169] The generated warning information is stored in the system database in a structured data format and simultaneously pushed to the central monitoring interface of the coal yard to realize the linkage release of graphic annotation and warning text.

[0170] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0171] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in the embodiments of this application are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0172] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0173] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0174] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0175] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0176] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0177] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0178] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0179] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A coal yard safety analysis system based on infrared imaging and point cloud data fusion, characterized in that: include: The data acquisition module collects infrared imaging data and three-dimensional point cloud data on the surface of the coal pile, performs multi-source information fusion processing on the infrared imaging data and the three-dimensional point cloud data, and generates an initial emissivity distribution map of the coal pile surface; The meshing module divides the coal pile surface into multiple local mesh areas based on the initial emissivity distribution map, and extracts the texture features and surface structure information of each local mesh area; The emission correction module performs local dynamic compensation on the initial emissivity distribution map based on texture features and surface structure information to generate a corrected emissivity distribution map; Establishing a mapping relationship between each local grid area in the initial emissivity distribution map and the corresponding texture features and surface structure information; In each local grid area, a spatial structure model is constructed based on the three-dimensional point cloud data; In each local grid area, an image texture feature model is constructed based on the infrared radiation intensity distribution image; The spatial structure model is jointly fitted with the image texture feature model to obtain the emissivity adjustment factor of each local grid area; Based on the emissivity adjustment factor of each local grid area, the original emissivity value is locally dynamically compensated to generate a corrected emissivity distribution map; The anomaly detection module identifies abnormal areas on the coal pile surface based on the corrected emissivity distribution map and marks them as potential risk areas; The diffusion assessment module analyzes the heat distribution characteristics in the potential risk area through the heat transfer diffusion model and evaluates the time dependence of heat diffusion; The fractal assessment module uses fractal analysis to assess the degree of heat distribution balance in potential risk areas; The risk warning module evaluates the safety risk level of the coal yard based on the time dependence of heat diffusion and the balance of heat distribution, and generates coal yard safety warning information.

2. The coal yard safety analysis system based on infrared imaging and point cloud data fusion according to claim 1 is characterized in that: Infrared imaging data and three-dimensional point cloud data are collected on the coal pile surface, and multi-source information fusion processing is performed on the infrared imaging data and the three-dimensional point cloud data to generate the initial emissivity distribution map of the coal pile surface, specifically: Infrared thermal imaging equipment is used to collect infrared imaging data of the coal pile surface to obtain an infrared radiation intensity distribution image; Use 3D laser scanning equipment to collect spatial position coordinate data and spatial surface structure feature data of the coal pile surface to obtain 3D point cloud data of the coal pile surface; Perform spatial coordinate alignment and time stamp alignment on the infrared radiation intensity distribution image and the three-dimensional point cloud data, and establish a mapping relationship between the infrared image pixel points and the point cloud coordinate points; Based on the mapping relationship between infrared image pixels and point cloud coordinates, the initial emissivity distribution map of the coal pile surface is generated.

3. The coal yard safety analysis system based on infrared imaging and point cloud data fusion according to claim 2 is characterized in that: Based on the initial emissivity distribution map, the coal pile surface is divided into multiple local grid areas, and the texture features and surface structure information of each local grid area are extracted. Specifically: The local spatial boundary of the coal pile surface is determined based on the initial emissivity distribution map, and the coal pile surface is divided into multiple local grid areas; For each local grid area, surface structure information is extracted based on 3D point cloud data; For each local grid area, texture features are extracted based on the infrared radiation intensity distribution image.

4. The coal yard safety analysis system based on infrared imaging and point cloud data fusion according to claim 3 is characterized in that: According to the corrected emissivity distribution map, abnormal areas on the coal pile surface are identified and marked as potential risk areas, specifically: Perform spatial neighborhood analysis on the corrected emissivity distribution map to determine the local statistical threshold of emissivity in each local grid area; Determine the abnormal deviation degree of the emissivity in each local grid area based on the local statistical threshold, and generate emissivity abnormal intensity data; Determine the spatial position boundary of the abnormal area on the surface of the coal pile based on the abnormal emissivity intensity data; The abnormal area on the coal pile surface is divided into spatial connectivity to determine the overall spatial range of the abnormal area; The overall spatial range of the abnormal area is marked to obtain the potential risk area.

5. The coal yard safety analysis system based on infrared imaging and point cloud data fusion according to claim 4 is characterized in that: The heat transfer diffusion model is used to analyze the heat distribution characteristics in the potential risk area and evaluate the time dependence of heat diffusion, specifically: Establish a heat transfer and diffusion model based on the overall spatial scope of the potential risk area; Conduct numerical simulation of heat diffusion in the potential risk area based on the heat transfer diffusion model to determine the heat distribution data of each local grid area in the potential risk area; Analyze the heat diffusion trend based on the heat distribution data of each local grid area in the potential risk area; According to the heat diffusion trend, the time dependence of heat diffusion in the potential risk area is evaluated to obtain the time characteristic data of heat diffusion.

6. The coal yard safety analysis system based on infrared imaging and point cloud data fusion according to claim 5 is characterized in that: The fractal analysis method is used to evaluate the degree of heat distribution balance in the potential risk area, specifically: Extract heat distribution data of each local grid area within the potential risk area; Mapping the heat distribution data of each local grid area within the potential risk area into a spatial heat distribution matrix; Multi-scale fractal dimension calculation is performed on the spatial heat distribution matrix to obtain the fractal dimension sequence of potential risk areas at different spatial scales; According to the changing trend of the fractal dimension sequence, the spatial balance of heat distribution in the potential risk area is evaluated, and the spatial characteristic data of heat diffusion are obtained.

7. The coal yard safety analysis system based on infrared imaging and point cloud data fusion according to claim 6 is characterized in that: Based on the time dependence of heat diffusion and the balance of heat distribution, the safety risk level of the coal yard is assessed and coal yard safety warning information is generated. Specifically: Based on the heat diffusion time characteristic data and heat diffusion space characteristic data, the feature vector is spliced to obtain the comprehensive risk assessment input data set. Based on the preset coal yard risk assessment model and combined with the comprehensive risk assessment input data set, the risk level assessment of the potential risk area is performed and the risk level value of the potential risk area is output; Classify the potential risk areas into risk levels according to the risk level values and generate a risk level map of the potential risk areas; Generate corresponding coal yard safety warning information based on the risk level map.

Images