Ore lumpiness distribution analysis method, device, equipment and medium

Abstract

The invention relates to the technical field of ore lumpiness distribution analysis. The ore lumpiness distribution analysis method comprises the following steps: acquiring image information in an ore field and depth information acquired by sensing equipment; the depth information and the preprocessed image information are combined, and three-dimensional space coordinate information of the ore is generated through a three-dimensional reconstruction algorithm; inputting the three-dimensional space coordinate information of the ore into the intelligent lumpiness prediction model to generate lumpiness distribution information of the ore; correcting the lumpiness distribution information of the ore by using a lumpiness size standard, and removing abnormal values in the lumpiness distribution information of the ore to obtain optimized lumpiness distribution information; according to the optimized lumpiness distribution information, boulder rate distribution information of the ore is calculated, and meanwhile, the optimized lumpiness distribution information and the boulder rate distribution information of the ore are displayed through a visual chart. The method has the effect of accurately obtaining the lumpiness distribution information of the ore.

Description

Technical Field

[0001] The present invention relates to the technical field of ore lump size distribution analysis, and in particular to a method, device, equipment and medium for ore lump size distribution analysis. Background Art

[0002] Currently, the identification and distribution analysis of ore lump sizes are important tasks in mine exploitation, blasting effect evaluation and subsequent processing. The lump size distribution of ore directly affects the efficiency of crushing, screening and transportation. Accurate lump size distribution information can not only guide the optimization of blasting technology, but also improve the efficiency of processing equipment and reduce energy consumption. Therefore, how to quickly and accurately obtain the lump size distribution information of ore has become a key research direction in the mining field. Summary of the Invention

[0003] In order to accurately obtain the lump size distribution information of ore, the present application provides a method, device, equipment and medium for ore lump size distribution analysis.

[0004] The first above-mentioned invention object of the present application is achieved by the following technical solutions: A method for ore lump size distribution analysis, the method for ore lump size distribution analysis includes: Obtain image information in the ore yard and depth information collected by a sensing device; Remove the noise of the image information through Gaussian filtering to obtain the image information after noise removal, perform binarization processing on the image information after noise removal to obtain binarized image information, and further perform morphological erosion and dilation operations on the binarized image information to obtain the preprocessed image information; Combine the depth information and the preprocessed image information, and generate three-dimensional spatial coordinate information of the ore through a three-dimensional reconstruction algorithm; input the three-dimensional spatial coordinate information of the ore into an intelligent lump size prediction model to generate the lump size distribution information of the ore; Calibrate the lump size distribution information of the ore using the standard of lump size, eliminate the outliers in the lump size distribution information of the ore, and obtain the optimized lump size distribution information; According to the optimized lump size distribution information, calculate the large lump rate distribution information of the ore, and simultaneously display the optimized lump size distribution information and the large lump rate distribution information of the ore through a visualization chart.

[0005] By adopting the above technical solutions, it is possible to accurately generate the three-dimensional spatial coordinate information of the ore through the combined processing of the image information and depth information in the ore yard, thereby effectively reflecting the geometric shape of the ore yard. Through Gaussian filtering, binarization, and morphological operations, image noise is eliminated and the ore boundary features are enhanced, making the preprocessed image information have higher clarity and accuracy, and ensuring the quality of subsequent three-dimensional reconstruction. Combining the three-dimensional reconstruction algorithm and depth information can effectively correct the coordinate deviation caused by camera distortion and projection error, and realize the generation of high-precision three-dimensional point cloud information, laying a foundation for the analysis of ore geometric characteristics. In addition, through the analysis and processing of the three-dimensional spatial coordinate information of the ore by the intelligent lump size prediction model, the lump size distribution information of the ore can be automatically generated, the geometric characteristics of the ore lumps can be quickly identified, and the size distribution ratio of the ore lumps can be accurately counted, effectively reducing the inaccuracy and time-consuming problems of manual measurement in the traditional method. Using the standard of lump size to correct and eliminate outliers ensures the accuracy and consistency of the lump size distribution information, thereby further optimizing the information analysis results. By calculating the large lump rate distribution information of the ore based on the optimized lump size distribution information, the volume ratio of the ore within the target particle size range can be clearly reflected, providing strong information support for the adjustment of blasting parameters and the optimization of transportation processes. Finally, by visualizing the lump size distribution information and large lump rate distribution information in charts, the intuitiveness and usability of information transmission are significantly improved, enabling technicians to quickly evaluate the lump size distribution characteristics of the ore yard.

[0006] In a preferred example of the present application, it can be further configured that: combining the depth information and the preprocessed image information, and generating the three-dimensional spatial coordinate information of the ore through a three-dimensional reconstruction algorithm, including: Extracting the edge pixel coordinate information and boundary contour information from the preprocessed image information, and generating the initial two-dimensional pixel coordinates according to the edge pixel coordinate information and the boundary contour information; Using the internal parameter matrix in the three-dimensional reconstruction algorithm to correct the initial two-dimensional pixel coordinates, and generating the corrected two-dimensional pixel coordinates; Combining the corrected two-dimensional pixel coordinates with the depth information by using the external parameter matrix in the three-dimensional reconstruction algorithm, and generating the preliminary three-dimensional spatial coordinate information; Performing outlier detection on the preliminary three-dimensional spatial coordinate information, removing the noise points and outliers, and generating the three-dimensional spatial coordinate information of the ore after completion.

[0007] By adopting the above technical solutions, it is possible to accurately extract the edge pixel coordinate information and boundary contour information from the preprocessed image information, effectively retain the geometric features of the ore area, and generate the initial two-dimensional pixel coordinates, providing reliable input information for subsequent three-dimensional reconstruction. By calibrating the two-dimensional pixel coordinates with the intrinsic matrix, the pixel offset caused by lens distortion and optical distortion is eliminated, ensuring that the calibrated two-dimensional pixel coordinates can more truly reflect the position of the ore in the image, thereby improving the accuracy of three-dimensional reconstruction. Combining the extrinsic matrix to map the calibrated two-dimensional pixel coordinates and depth information can efficiently generate preliminary three-dimensional spatial coordinate information, ensuring that the spatial position of each pixel point is accurately mapped into the global coordinate system, effectively reflecting the geometric structure of the ore pile. By detecting outliers in the preliminary three-dimensional spatial coordinate information and removing noise points and outliers, the quality of the three-dimensional point cloud information can be significantly improved, avoiding the influence of noise or outliers on subsequent information analysis and processing, thereby generating accurate and continuous three-dimensional spatial coordinate information of the ore.

[0008] In a preferred example of the present application, it can be further configured that: calibrating the initial two-dimensional pixel coordinates with the intrinsic matrix in the three-dimensional reconstruction algorithm to generate calibrated two-dimensional pixel coordinates includes: Performing optical distortion calibration on the two-dimensional pixel coordinates based on the parameters in the intrinsic matrix, and eliminating the influence of lens distortion through the following formula to generate calibrated two-dimensional pixel coordinates: where u is the value of the two-dimensional pixel coordinate in the horizontal direction, v is the value of the two-dimensional pixel coordinate in the vertical direction, x′ is the value of the calibrated two-dimensional pixel coordinate in the horizontal direction, y′ is the value of the calibrated two-dimensional pixel coordinate in the vertical direction, and f x is the horizontal focal length parameter of the intrinsic matrix, f y is the vertical focal length parameter of the intrinsic matrix, c x is the horizontal coordinate of the camera principal point, and c y is the vertical coordinate of the camera principal point.

[0009] By adopting the above technical solutions, it is possible to accurately perform optical distortion calibration on the two-dimensional pixel coordinates based on the parameters in the intrinsic matrix. The pixel offset and geometric deformation problems caused by lens distortion are eliminated through formula calibration, thereby effectively restoring the true spatial position of the pixel points in the image. By calibrating the horizontal focal length parameter and the vertical focal length parameter, the problem of image scale distortion is adjusted, making the calibrated two-dimensional pixel coordinates more accurate in the horizontal and vertical directions. At the same time, the offset of the pixel points is compensated by using the horizontal and vertical coordinates of the camera principal point, effectively aligning the mapping position of the pixel points at the center of the image, thereby ensuring the symmetry and accuracy of the pixel distribution.

[0010] In a preferred example, the present application can be further configured as follows: combining the corrected two-dimensional pixel coordinates with the depth information by using the external parameter matrix in the three-dimensional reconstruction algorithm to generate preliminary three-dimensional spatial coordinate information, including: Based on the rotation matrix and translation vector in the external parameter matrix, combining the corrected two-dimensional pixel coordinates with the depth information, and calculating the preliminary three-dimensional spatial coordinate information standard through the following formula: Wherein, X is the value of the three-dimensional spatial coordinate in the horizontal direction, Y is the value of the three-dimensional spatial coordinate in the vertical direction, Z is the depth information, R is the rotation matrix in the external parameter matrix, T is the translation vector in the external parameter matrix, and x' and y' are the horizontal direction value and vertical direction value of the corrected two-dimensional pixel coordinates, respectively.

[0011] By adopting the above technical solution, it is possible to efficiently combine the corrected two-dimensional pixel coordinates with the depth information based on the rotation matrix and translation vector in the external parameter matrix, and accurately calculate the preliminary three-dimensional spatial coordinate information through the formula. The rotation matrix can accurately describe the rotation relationship between the camera coordinate system and the global coordinate system, so as to perform direction conversion and alignment on the two-dimensional pixel coordinates. The translation vector can compensate for the displacement of the camera coordinate system relative to the global coordinate system, so that the coordinate points are mapped from the camera perspective to the actual spatial position. The combination of the depth information ensures that the spatial position of each pixel point has an actual three-dimensional depth value, and can fully reflect the geometric structure characteristics of the ore yard.

[0012] In a preferred example, the present application can be further configured as follows: inputting the three-dimensional spatial coordinate information of the ore into an intelligent block size prediction model to generate the block size distribution information of the ore, including: Extracting the feature information of the ore by performing feature extraction on the three-dimensional spatial coordinate information of the ore; Classifying the ore according to the preset ore classification standard based on the ore feature information, determining the size range of each ore, and generating an ore classification result; Performing statistical analysis on the ore size ranges in the ore classification result, calculating the quantity ratio of the ore according to different size ranges, and generating the block size distribution information of the ore.

[0013] By adopting the above technical solution, it is possible to accurately extract the three-dimensional spatial coordinate information of the ore, effectively obtain the geometric feature information of the ore, including features such as volume, surface area, circumscribed sphere diameter, and length-width ratio, etc., providing comprehensive and accurate basic information for subsequent classification and analysis. By combining the preset ore classification criteria, the ore can be efficiently classified according to the size range, and the classification result can be automatically generated, avoiding the errors and low efficiency problems existing in the traditional manual classification process, and ensuring the accuracy and consistency of ore classification.

[0014] In a preferred example of the present application, it can be further configured as follows: using the standard of the lump size to correct the lump size distribution information of the ore, removing the outliers in the lump size distribution information of the ore, and obtaining the optimized lump size distribution information, including: According to the preset lump size standard, compare the ore sizes in the lump size distribution information of the ore one by one. During the comparison process, identify the outliers that exceed the range of the lump size standard; Remove the outliers from the lump size distribution information of the ore to obtain the ore size information after removing the outliers. According to the ore size information after removing the outliers, recalculate the lump size distribution information of the ore to obtain the recalculated lump size distribution information of the ore; Statistically analyze the recalculated lump size distribution information of the ore, re-divide the proportion of ores in different lump size ranges according to the corrected ore size range, and generate the optimized lump size distribution information of the ore.

[0015] By adopting the above technical solution, it is possible to compare the ore sizes in the lump size distribution information of the ore one by one according to the preset lump size standard, effectively identify and remove the outliers that exceed the range of the lump size standard, thereby avoiding the interference of abnormal information on the accuracy of the lump size distribution information. After removing the outliers, recalculating the lump size distribution information of the ore can ensure the integrity and authenticity of the information, making the calculation result more in line with the actual situation. In addition, statistically analyzing the recalculated lump size distribution information of the ore and re-dividing the proportion of different lump size ranges according to the corrected ore size range can comprehensively reflect the true distribution of the ore lump size, and the generated optimized lump size distribution information has higher accuracy and consistency.

[0016] In a preferred example of the present application, it can be further configured as follows: calculating the large lump rate distribution information of the ore according to the optimized lump size distribution information, including: Obtain the volume proportion information of each particle size range of the ore from the optimized lump size distribution information; Obtain the target particle size threshold, compare the volume proportion information of each particle size range of the ore with the target particle size threshold, and generate the volume proportion information of the ore particle size greater than the target particle size threshold; Normalize the volume proportion information of the ore particle size, calculate the proportion of the volume of the ore greater than the target particle size threshold in the total ore volume, and obtain the large block rate distribution information of the ore.

[0017] By adopting the above technical solution, it is possible to efficiently extract the volume proportion information of each particle size range of the ore from the optimized block size distribution information, providing an accurate information basis for subsequent analysis. By setting the target particle size threshold and comparing the extracted particle size volume proportion information with the target particle size threshold one by one, it is possible to accurately identify and screen out the ore volume information greater than the target particle size range, thereby generating the volume proportion information of the ore particle size greater than the target particle size threshold, and achieving precise classification of the ore within the target particle size range. In addition, by normalizing the volume proportion information of the ore particle size, the influence brought by different information scales can be effectively eliminated, making the volume proportion result more standardized and comparable. Further calculate the proportion of the volume of the ore greater than the target particle size threshold in the total ore volume to generate the large block rate distribution information of the ore. The large block rate distribution information can intuitively reflect the proportion of the ore above the target particle size in the overall ore pile, providing important information support for the evaluation of the blasting effect of the mine, the selection of transportation equipment, and the optimization of subsequent processing technology.

[0018] The above second inventive object of the present application is achieved by the following technical solutions: An ore block size distribution analysis device, the ore block size distribution analysis device includes: An ore image and depth information acquisition module, used to acquire the image information in the ore field and the depth information collected by the sensing device; an image preprocessing module, used to remove the noise of the image information through Gaussian filtering to obtain the noise-removed image information, perform binary processing on the noise-removed image information to obtain binary image information, and further perform morphological erosion and dilation operations on the binary image information to obtain the preprocessed image information; A three-dimensional space coordinate generation module, used to combine the depth information and the preprocessed image information to generate the three-dimensional space coordinate information of the ore through a three-dimensional reconstruction algorithm; A block size distribution generation module, used to input the three-dimensional space coordinate information of the ore into an intelligent block size prediction model to generate the block size distribution information of the ore; A block size distribution correction module, used to correct the block size distribution information of the ore using the standard of the block size, eliminate the outliers in the block size distribution information of the ore, and obtain the optimized block size distribution information; The large block rate calculation and visualization module is used to calculate the large block rate distribution information of the ore according to the optimized block size distribution information, and at the same time display the optimized block size distribution information and the large block rate distribution information of the ore through a visualization chart.

[0019] By adopting the above technical solution, it is possible to accurately generate the three-dimensional space coordinate information of the ore through the combined processing of the image information and depth information in the ore yard, so as to effectively reflect the geometric shape of the ore yard. Through Gaussian filtering, binarization and morphological operations, image noise is eliminated and the ore boundary features are enhanced, making the preprocessed image information have higher clarity and accuracy, and ensuring the quality of subsequent three-dimensional reconstruction. Combining the three-dimensional reconstruction algorithm and depth information can effectively correct the coordinate deviation caused by camera distortion and projection error, and realize the generation of high-precision three-dimensional point cloud information, laying a foundation for the analysis of ore geometric characteristics. In addition, through the analysis and processing of the three-dimensional space coordinate information of the ore by the intelligent block size prediction model, the block size distribution information of the ore can be automatically generated, the geometric characteristics of the ore blocks can be quickly identified, and the size distribution ratio of the ore blocks can be accurately counted, effectively reducing the inaccuracy and time-consuming problems of manual measurement in the traditional method. Using the standard of block size to correct and eliminate outliers ensures the accuracy and consistency of the block size distribution information, thus further optimizing the information analysis results. By calculating the large block rate distribution information of the ore based on the optimized block size distribution information, the volume ratio of the ore in the target particle size range can be clearly reflected, providing strong information support for the adjustment of blasting parameters and the optimization of transportation technology. Finally, by displaying the block size distribution information and the large block rate distribution information through a visualization chart, the intuitiveness and usability of information transmission are significantly improved, enabling technicians to quickly evaluate the block size distribution characteristics of the ore yard. The above-mentioned third object of the present application is achieved by the following technical solution: A device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned ore block size distribution analysis method are realized.

[0020] The above-mentioned fourth object of the present application is achieved by the following technical solution: A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned ore block size distribution analysis method are realized.

[0021] In summary, the present application includes at least one of the following beneficial technical effects: 1. It can accurately generate the three-dimensional spatial coordinate information of ore by combining and processing the image information and depth information in the ore yard, thus effectively reflecting the geometric shape of the ore yard. Through Gaussian filtering, binarization, and morphological operations, image noise is eliminated and the ore boundary features are enhanced, making the preprocessed image information have higher clarity and accuracy, and ensuring the quality of subsequent three-dimensional reconstruction. By combining the three-dimensional reconstruction algorithm and depth information, the coordinate deviation caused by camera distortion and projection error can be effectively corrected, realizing the generation of high-precision three-dimensional point cloud information, laying a foundation for the analysis of ore geometric characteristics. In addition, through the analysis and processing of the three-dimensional spatial coordinate information of ore by the intelligent lump size prediction model, the lump size distribution information of ore can be automatically generated, the geometric characteristics of ore lumps can be quickly identified, and the size distribution ratio of ore lumps can be accurately counted, effectively reducing the inaccuracy and time-consuming problems of manual measurement in traditional methods. By correcting and removing outliers using the standard of lump size, the accuracy and consistency of the lump size distribution information are ensured, thus further optimizing the information analysis results. By calculating the large lump rate distribution information of ore based on the optimized lump size distribution information, the volume ratio of ore within the target particle size range can be clearly reflected, providing strong information support for blasting parameter adjustment and transportation process optimization. Finally, by visualizing the lump size distribution information and large lump rate distribution information through charts, the intuitiveness and usability of information transmission are significantly improved, enabling technicians to quickly evaluate the lump size distribution characteristics of the ore yard; 2. It can accurately extract the features of the three-dimensional spatial coordinate information of ore, effectively obtain the geometric feature information of ore, including features such as volume, surface area, circumscribed sphere diameter, and length-width ratio, providing comprehensive and accurate basic information for subsequent classification and analysis. By combining the preset ore classification criteria, ores can be efficiently classified according to the size range, and the classification results can be automatically generated, avoiding the errors and low efficiency problems existing in the traditional manual classification process, and ensuring the accuracy and consistency of ore classification; 3. It can compare the ore sizes in the ore lump size distribution information one by one according to the preset lump size standard, effectively identify and remove the outliers outside the lump size standard range, thereby avoiding the interference of abnormal information on the accuracy of the lump size distribution information. After removing the outliers and recalculating the ore lump size distribution information, the integrity and authenticity of the information can be ensured, making the calculation results more in line with the actual situation. In addition, by statistically analyzing the recalculated ore lump size distribution information and re-dividing the proportion of different lump size ranges according to the corrected ore size range, the true distribution of ore lump sizes can be comprehensively reflected, and the optimized lump size distribution information generated has higher accuracy and consistency. Description of the Drawings

[0022] Figure 1 is a flowchart of the ore lump size distribution analysis method in an embodiment of the present application; Figure 2 It is the implementation flowchart of step S30 in the ore lump size distribution analysis method in an embodiment of the present application; Figure 3 It is the implementation flowchart of step S302 in the ore lump size distribution analysis method in an embodiment of the present application; Figure 4 It is the implementation flowchart of step S303 in the ore lump size distribution analysis method in an embodiment of the present application; Figure 5 It is the implementation flowchart of step S40 in the ore lump size distribution analysis method in an embodiment of the present application; Figure 6 It is the implementation flowchart of step S50 in the ore lump size distribution analysis method in an embodiment of the present application; Figure 7 It is the implementation flowchart of step S60 in the ore lump size distribution analysis method in an embodiment of the present application; Figure 8 It is a principle block diagram of an ore lump size distribution analysis device in an embodiment of the present application; Figure 9 It is a schematic diagram of the equipment in an embodiment of the present application. Detailed implementation manners

[0023] The present application will be further described in detail below with reference to the accompanying drawings.

[0024] In one embodiment, as Figure 1 shown, the present application discloses an ore lump size distribution analysis method, which specifically includes the following steps: S10: Obtain the image information in the ore yard and the depth information collected by the sensing device.

[0025] Specifically, when obtaining the image information in the ore yard, a high-resolution camera device is used to take pictures of the ore yard from multiple angles and positions, and the shooting height and focal length of the camera device are adjusted to cover the entire ore yard, ensuring that the picture is clear and the details of the ore surface are visible. The acquisition of the depth information is achieved by installing a lidar or a structured light depth sensor. The distance from each point on the surface of the ore pile to the sensor is measured by the lidar, and the depth information is recorded by the time-of-flight method or the triangulation method. At the same time, the information of the depth sensor is aligned with the image information of the camera device point by point in pixels, and the depth information and the image information are coordinate-mapped through the internal parameter matrix after camera calibration, so as to obtain the image information and the depth information of the ore yard at the same time.

[0026] S20: Remove the noise in the image information through Gaussian filtering to obtain the image information after noise removal. Perform binarization processing on the image information after noise removal to obtain binarized image information. Further perform morphological erosion and dilation operations on the binarized image information to obtain the preprocessed image information.

[0027] Specifically, when removing the noise in the image information through Gaussian filtering, first select an appropriate two-dimensional Gaussian filter kernel, the size of which is set according to the image resolution and noise intensity. The Gaussian filter kernel is implemented by constructing a two-dimensional Gaussian function. The formula of the Gaussian function is: where x and y are the horizontal and vertical distances between the current pixel and the central pixel of the filter respectively, and σ is the standard deviation of the Gaussian distribution, which is used to control the smoothness of the filter. When the filter performs a convolution operation on the image, the value of each pixel in the image is weighted and averaged with its neighboring pixels according to the weights distributed by the Gaussian function, and the calculated value replaces the value of the central pixel, so as to effectively remove the interference of random noise while retaining the edge details in the image, obtaining the image information after noise removal. When performing binarization processing on the image information after noise removal, first map the pixel values of the grayscale image to a fixed range (such as 0 to 255), and then set a binarization threshold, which can be determined by a fixed threshold or dynamically calculated by the Otsu algorithm. For the part where the pixel value is greater than the threshold, it is assigned a value of 1 (foreground pixel), and for the part where the pixel value is less than or equal to the threshold, it is assigned a value of 0 (background pixel), thus generating binarized image information. The result of binarization retains the main contour information of the ore area while removing the irrelevant background information. When further performing morphological erosion and dilation operations on the binarized image information, the erosion operation selects a structuring element (such as a 3×3 or 5×5 square structure), and performs a logical AND operation on the central pixel of the structuring element and its neighboring pixels. Only when all the pixels within the coverage of the structuring element are foreground pixels, the central pixel is retained as a foreground pixel, otherwise it is set as a background pixel, thus realizing the removal of isolated noise points and the contraction of the boundary in the binarized image. The dilation operation performs a logical OR operation through the same structuring element. When any one of the pixels within the coverage of the structuring element is a foreground pixel, the central pixel is a foreground pixel, thus filling the fractures or voids in the ore area and enhancing the coherence of the ore boundary. After erosion and dilation processing, a more complete and coherent ore area is generated, completing the preprocessing of the image information.

[0028] S30: Combine the depth information and the preprocessed image information to generate the three-dimensional spatial coordinate information of the ore through a three-dimensional reconstruction algorithm.

[0029] Specifically, when combining depth information and pre - processed image information, first, extract the edge pixel coordinate information and boundary contour information of the ore surface from the pre - processed image. The extraction of edge pixel coordinate information can be completed by calculating the pixel gradient of the image. By setting a threshold to detect edge pixel points and extract the connected boundary contour information, the basic two - dimensional shape of the ore block is determined. Take the two - dimensional coordinates of these edge pixel points as the initial image feature information and perform one - by - one corresponding matching with the depth value of each pixel point in the depth information. Through a three - dimensional reconstruction algorithm, use the internal parameter matrix of the imaging device to correct the extracted two - dimensional pixel coordinates. The internal parameter matrix contains the optical characteristic parameters of the imaging device. By correcting the two - dimensional pixel coordinates, eliminate the pixel offset caused by optical distortion to obtain the corrected pixel coordinates. After combining the corrected pixel coordinates with the depth information, initially generate the three - dimensional coordinate information in the camera coordinate system. Further, complete the coordinate transformation through the external parameter matrix in the three - dimensional reconstruction algorithm. The external parameter matrix contains a rotation matrix and a translation vector. The rotation matrix is used to describe the rotation relationship between the camera coordinate system and the global coordinate system, and the translation vector is used to describe the displacement relationship from the camera coordinate system to the global coordinate system. Through the external parameter matrix, transform the three - dimensional coordinate information in the camera coordinate system into the global coordinate system, and finally generate the three - dimensional space coordinate information in the global coordinate system.

[0030] S40: Input the three - dimensional space coordinate information of the ore into the intelligent lumpiness prediction model to generate the lumpiness distribution information of the ore.

[0031] Specifically, by taking the three - dimensional space coordinate information as the input information, first extract the spatial geometric features of each ore block, including parameters such as volume, surface area, and circum - sphere diameter. The volume is determined by analyzing the spatial distribution relationship of all points in the three - dimensional point cloud information. The surface area is obtained by calculating the connecting surfaces of the points in the point cloud. The circum - sphere diameter is completed by fitting the diameter value of the minimum circum - sphere of the ore block in the three - dimensional coordinate information. The extracted geometric feature information can be used to describe the shape and size of the ore block. Subsequently, according to the preset ore particle size interval standard, classify the ore in combination with the three - dimensional spatial geometric features. By comparing the geometric parameters of each ore block with the standard particle size range, divide the ore blocks into the corresponding particle size intervals. After classification, count the number, volume ratio, and volume percentage of the ore blocks in each particle size interval in the total ore pile. Take these classification results as the basic information of the ore lumpiness distribution information. After completing classification and statistics, further process the classified information. By cumulatively calculating the number and volume ratio of the ore in each particle size interval, generate a complete lumpiness distribution information table. The lumpiness distribution information table records the particle size distribution of the ore blocks and the ratio information of each particle size range, and finally output the lumpiness distribution information of the ore.

[0032] S50: Calibrate the lump size distribution information of the ore according to the standard of lump size, eliminate the outliers in the lump size distribution information of the ore, and obtain the optimized lump size distribution information.

[0033] Specifically, according to the lump size standard, compare each particle size range in the lump size distribution information of the ore one by one, identify the outliers by detecting the ore information outside the lump size range, and process the identified outliers, including recalculating its particle size or directly removing it from the information set. After removing the outliers, recalculate the proportion of the ore volume in each particle size range, and record the updated statistical information as the optimized lump size distribution information. The optimized information covers all the ore particle size ranges that meet the lump size standard, and reduces the error introduced by the abnormal points through the calibrated information.

[0034] S60: Calculate the large lump rate distribution information of the ore according to the optimized lump size distribution information, and at the same time display the optimized lump size distribution information and the large lump rate distribution information of the ore through a visualization chart.

[0035] Specifically, according to the optimized lump size distribution information, first set a target particle size threshold, which can be determined according to the input requirements of the ore processing equipment or the transportation standard. The target particle size threshold is used to distinguish large lump ore and small lump ore. By screening the volume information of each particle size range in the optimized lump size distribution information one by one, select the ore information with a particle size greater than the target threshold, extract the proportion information of the ore volume greater than the target particle size threshold, and compare it with the total ore volume to calculate the proportion of large lump ore in the total ore pile volume, and record this proportion as the large lump rate of the ore. After calculating the large lump rate distribution information, organize the optimized lump size distribution information and the large lump rate distribution information into visualization information, and display the results by constructing a histogram, a pie chart or a distribution curve diagram, etc. The histogram is used to display the proportion distribution of the ore in different particle size ranges, the pie chart is used to display the proportion of the total volume occupied by the ore with different particle sizes, and the distribution curve is used to display the change trend of the ore lump size and the fluctuation of the large lump rate. Display these charts in a clear and intuitive way to facilitate users to quickly understand the lump size distribution characteristics of the ore and the statistical results of the large lump rate.

[0036] In one embodiment, as Figure 2 shown, in step S30, that is, combining the depth information and the preprocessed image information, generate the three-dimensional spatial coordinate information of the ore through a three-dimensional reconstruction algorithm, including: S301: Extract the edge pixel coordinate information and the boundary contour information from the preprocessed image information, and generate the initial two-dimensional pixel coordinates according to the edge pixel coordinate information and the boundary contour information.

[0037] Specifically, when extracting the edge pixel coordinate information from the preprocessed image information, it can be completed by using an edge detection algorithm. When using the Canny edge detection algorithm, first, the image is grayscaled, and then by calculating the gradient intensity and direction of the pixels, the pixels with drastic gray changes in the image are marked as edge pixels. Then, non-maximum suppression is applied to remove false edge points, and only the pixels with local maximum values in the gradient direction are retained as valid edge pixel points. Finally, through double-threshold detection and edge tracking connection, complete edge information is formed, and the boundary contour information of the ore block is generated based on the extracted edge pixel points. The contour information can be obtained through the connectivity analysis of the pixel points, and the coordinates of these edge pixel points are organized into initial two-dimensional pixel coordinates.

[0038] S302: Use the internal parameter matrix in the three-dimensional reconstruction algorithm to correct the initial two-dimensional pixel coordinates and generate the corrected two-dimensional pixel coordinates.

[0039] Specifically, when using the internal parameter matrix to correct the initial two-dimensional pixel coordinates, first, according to the focal length parameter and the coordinates of the optical center point in the camera internal parameter matrix, distortion correction is performed on the initial two-dimensional pixel coordinates. By analyzing the optical distortion situation in the image, the positions of each pixel point in the horizontal and vertical directions are adjusted, and the offset pixel points are remapped to the corrected coordinate positions. The internal parameter correction can not only eliminate the image distortion caused by lens distortion but also adjust the scale and distribution of the initial two-dimensional pixel coordinates, making the corrected two-dimensional pixel coordinates more in line with the optical characteristics of the camera model and providing accurate input for subsequent depth information combination and three-dimensional reconstruction.

[0040] S303: Use the external parameter matrix in the three-dimensional reconstruction algorithm to combine the corrected two-dimensional pixel coordinates with the depth information and generate preliminary three-dimensional space coordinate information.

[0041] Specifically, when using the external parameter matrix in the three-dimensional reconstruction algorithm to combine the corrected two-dimensional pixel coordinates with the depth information, first, the corrected two-dimensional pixel coordinates are matched with the depth information one by one. The depth value of each pixel point represents its actual distance from the camera optical center. Through the combination of the depth information, each pixel point is mapped from the two-dimensional image plane to the three-dimensional space, and preliminary three-dimensional coordinate information in the camera coordinate system is generated. Subsequently, the coordinate system conversion is completed through the external parameter matrix. The external parameter matrix includes a rotation matrix and a translation vector. The rotation matrix is used to rotate the three-dimensional coordinate points in the camera coordinate system to the direction of the global coordinate system, and the translation vector is used to translate the rotated three-dimensional coordinate points to the reference position of the global coordinate system. At the same time, to ensure the continuity of the information, the depth values of the possibly discontinuous depth points are smoothed by the weighted average method. After the information correction, all point coordinates are uniformly stored as point cloud information in the global coordinate system to generate preliminary three-dimensional space coordinate information.

[0042] S304: Perform outlier detection on the preliminary three-dimensional spatial coordinate information, remove noise points and outliers, and generate the three-dimensional spatial coordinate information of the ore after completion.

[0043] Specifically, when performing outlier detection on the preliminary three-dimensional spatial coordinate information, first perform neighborhood analysis on each point in the point cloud information, statistically calculate the average distance and distribution characteristics between each point and its neighborhood points, identify outliers according to the set distance threshold. When the average distance of a certain point from its neighborhood is significantly higher than the threshold, this point is marked as a noise point or an outlier. Subsequently, further verify the spatial distribution of each point through clustering analysis, apply a density clustering algorithm (such as DBSCAN) to cluster the point cloud, divide the regions with close connections in the point cloud into valid point clusters, and the isolated points or small point clusters are marked as outliers. For the identified noise points and outliers, they are processed by direct removal or smoothing correction. When removing, the noise points are directly removed from the point cloud information to avoid interference with the overall distribution of the three-dimensional coordinates. When correcting, the position of the noise point is re-estimated according to the neighborhood distribution, and the new coordinate value is calculated by weighted averaging the positions of the neighborhood points to further ensure the continuity and accuracy of the point cloud information. After completing the outlier processing, all the remaining points are reorganized, and the valid three-dimensional point cloud information is sorted and stored according to the global coordinate system, and finally the three-dimensional spatial coordinate information of the ore is generated.

[0044] In one embodiment, as Figure 3 shown, in step S302, that is, use the internal parameter matrix in the three-dimensional reconstruction algorithm to correct the initial two-dimensional pixel coordinates to generate the corrected two-dimensional pixel coordinates, including: S3021: Perform optical distortion correction on the two-dimensional pixel coordinates based on the parameters in the internal parameter matrix, eliminate the influence of lens distortion through the following formula, and generate the corrected two-dimensional pixel coordinates: where u is the value of the two-dimensional pixel coordinate in the horizontal direction, v is the value of the two-dimensional pixel coordinate in the vertical direction, x′ is the value of the corrected two-dimensional pixel coordinate in the horizontal direction, y′ is the value of the corrected two-dimensional pixel coordinate in the vertical direction, f x is the horizontal focal length parameter of the internal parameter matrix, f y is the vertical focal length parameter of the internal parameter matrix, c x is the horizontal coordinate of the camera principal point, c y is the vertical coordinate of the camera principal point.

[0045] Specifically, during optical distortion correction, first, the horizontal focal length parameter, vertical focal length parameter, and the horizontal and vertical coordinates of the optical center point are extracted from the camera's internal parameter matrix. Subsequently, according to the camera's perspective projection model, geometric mapping calculations are performed on the two-dimensional pixel coordinates. By adjusting the coordinate positions of the pixel points, the distortion effects caused by lens distortion are corrected. Especially for the pixel points in the edge region of the image, their offset is usually large. Each pixel point's position is corrected point by point through the correction formula, and its true pixel position is mapped back to the undistorted pixel coordinate system. During the correction process, considering the radial distortion and tangential distortion characteristics of the lens, the correction amounts for the radial offset and tangential offset of each pixel are calculated respectively according to the correction formula, and then the position of the initial two-dimensional pixel coordinates is adjusted so that the position of each pixel point meets the requirements of the ideal projection model. The generated two-dimensional pixel coordinates after correction retain the actual geometric characteristics of the image and at the same time eliminate the geometric deformation caused by optical distortion, and finally complete the generation of the corrected two-dimensional pixel coordinates.

[0046] For example: during optical distortion correction, first, the parameters of the imaging device are extracted from the internal parameter matrix, including the horizontal focal length parameter f x = 800 pixels, the vertical focal length parameter f y = 800 pixels, the horizontal coordinate c x = 640 of the optical center point, and the vertical coordinate c y = 360. Then, the initial two-dimensional pixel coordinates are substituted into the correction formula for point-by-point correction. Assume the initial two-dimensional pixel coordinate values are u = 700 and v = 400. According to the calculation of the correction formula, first, the corrected horizontal pixel coordinate is calculated, and then the corrected vertical pixel coordinate is calculated. After calculation, the corrected two-dimensional pixel coordinates are x' = 0.075 and y' = 0.05.

[0047] In one embodiment, as Figure 4 shown, in step S303, that is, the corrected two-dimensional pixel coordinates are combined with the depth information using the external parameter matrix in the three-dimensional reconstruction algorithm to generate preliminary three-dimensional spatial coordinate information, including: S3031: Based on the rotation matrix and translation vector in the external parameter matrix, the corrected two-dimensional pixel coordinates are combined with the depth information, and the preliminary three-dimensional spatial coordinate information standard is calculated through the following formula: where X is the value of the three-dimensional spatial coordinate in the horizontal direction, Y is the value of the three-dimensional spatial coordinate in the vertical direction, Z is the depth information, R is the rotation matrix in the external parameter matrix, T is the translation vector in the external parameter matrix, and x' and y' are the horizontal direction value and vertical direction value of the corrected two-dimensional pixel coordinates respectively.

[0048] Specifically, when combining the corrected two-dimensional pixel coordinates with depth information using the rotation matrix and translation vector in the extrinsic parameter matrix, first, the depth value of the pixel point is determined through the depth information. The depth value is used as the Z-axis value of the pixel point in the camera coordinate system, and the corrected two-dimensional pixel coordinates are used to calculate its X-axis and Y-axis values in the camera coordinate system. Subsequently, through matrix operations on the three-dimensional coordinate information in the camera coordinate system and the rotation matrix and translation vector in the extrinsic parameter matrix, the conversion from the camera coordinate system to the global coordinate system is completed. The rotation matrix is used to describe the rotation relationship between the camera coordinate system and the global coordinate system. The three-dimensional coordinate values of the pixel point are rotated according to the calculation rules of the rotation matrix, and the original three-dimensional coordinate values are mapped to the direction of the global coordinate system. The translation vector is used to describe the offset position of the origin of the camera coordinate system relative to the global coordinate system. By adding the displacement of the translation vector to the rotated three-dimensional coordinate values, the coordinate position of the pixel point is converted from the reference frame of the camera coordinate system to the reference frame of the global coordinate system. After rotation and translation operations, preliminary three-dimensional spatial coordinate information is generated.

[0049] For example, when combining the corrected two-dimensional pixel coordinates with depth information using the rotation matrix and translation vector in the extrinsic parameter matrix, assume that the value of a corrected two-dimensional pixel coordinate is x′ = 0.1, y′ = 0.2, and the depth value corresponding to the depth information is Z = 2.5 meters. The rotation matrix R in the extrinsic parameter matrix is as follows: The translation vector T is as follows: The corrected two-dimensional pixel coordinates are converted into three-dimensional coordinate values in the camera coordinate system through the depth value, and the calculation results are: Next, the three-dimensional coordinates of the camera coordinate system are converted to the global coordinate system through the rotation matrix and translation vector, and the calculation is as follows: The values of the three-dimensional spatial coordinates in the global coordinate system are X = 0.75, Y = 1.0, and Z = 3.5.

[0050] In one embodiment, as Figure 5 shown, in step S40, the three-dimensional spatial coordinate information of the ore is input into the intelligent block size prediction model to generate the block size distribution information of the ore, including: S401: Extract features from the three-dimensional spatial coordinate information of the ore to extract ore feature information.

[0051] Specifically, when extracting the feature of the three-dimensional spatial coordinate information of the ore, first, by analyzing the spatial distribution of points in the three-dimensional point cloud information, the geometric features of each ore block are extracted, including the volume, surface area, circumscribed sphere diameter, and shape feature of the ore block. The extraction of the volume is realized by constructing the convex hull algorithm of the point cloud. The three-dimensional point cloud of the ore is wrapped in the smallest convex hull and the volume value it contains is calculated. The surface area is completed by extracting the surface mesh structure of each ore block in the point cloud and summing them up. The circumscribed sphere diameter is calculated by fitting the smallest enclosing sphere to the positions of all points in the point cloud and then calculating the diameter value of the sphere. The shape feature is extracted by analyzing the aspect ratio and surface smoothness of the point cloud. These geometric features are sorted into ore feature information.

[0052] S402: According to the ore feature information, use the preset ore classification standard to classify the ore, determine the size range of each ore, and generate the ore classification result.

[0053] Specifically, when classifying the ore according to the ore feature information, first, the feature information of each ore block is compared with the preset ore classification standard one by one. The classification standard includes the interval range of particle size. For example, the ore is divided into small blocks (particle size less than 50 mm), medium blocks (particle size between 50 mm and 200 mm), and large blocks (particle size greater than 200 mm). By matching the feature values such as the circumscribed sphere diameter or the aspect ratio of the point cloud of the ore block with the classification standard, the size range to which each ore block belongs is determined. Subsequently, the ore blocks are classified according to their respective categories to generate the classification result. The classification result includes the number of ore blocks in each size range and the corresponding geometric features. The classification process is completed using an efficient search algorithm to ensure the accurate and rapid classification of the ore blocks.

[0054] Furthermore, the search algorithm refers to the binary search algorithm. The binary search algorithm is used to sort the interval range of the classification standard in ascending order and then use the binary method to search for the feature values of the ore blocks. Each time the classification interval range is divided into two parts to determine the interval position where the feature value belongs, so as to quickly locate the classification to which the ore block belongs.

[0055] S403: Conduct a statistical analysis on the ore size range in the ore classification result, calculate the quantity ratio of the ore according to different size ranges, and generate the lump size distribution information of the ore.

[0056] Specifically, when statistically analyzing the ore size range in the ore classification results, first count the number of ore blocks in each size range, and calculate the proportion of the ore volume in this size range to the total ore volume. During the statistics, sum up the ore blocks one by one in each classification result, and accumulate to obtain the ore quantity and total volume of each classification. Then, based on the total number and total volume of all ore blocks in the classification results, calculate the quantity ratio and volume ratio of each size range, and organize this ratio information into the lump size distribution information, which records the distribution ratios of the ore quantity and volume in different particle size ranges.

[0057] In one embodiment, as Figure 6 shown, in step S50, that is, use the standard of lump size to correct the lump size distribution information of the ore, and eliminate the outliers in the lump size distribution information of the ore to obtain the optimized lump size distribution information, including: S501: According to the preset lump size standard, compare the ore sizes in the lump size distribution information of the ore one by one. During the comparison process, identify the outliers that exceed the range of the lump size standard.

[0058] Specifically, when comparing the ore sizes in the ore lump size distribution information one by one according to the preset lump size standard, first set the upper and lower limit ranges of the ore particle size according to the lump size standard. For example, the standard stipulates that the small lump range is a particle size less than 50 mm, the medium lump range is 50 - 200 mm, and the large lump range is greater than 200 mm. Then compare the particle size characteristic value (such as the circumscribed sphere diameter or the length-width-height ratio) of each ore block with the preset classification interval to check whether there is ore information with a particle size characteristic value exceeding the above standard range. When the particle size value of an ore block is higher than the maximum particle size value or lower than the minimum particle size value, mark this ore block as an outlier.

[0059] S502: Eliminate the outliers from the lump size distribution information of the ore to obtain the ore size information after eliminating the outliers. According to the ore size information after eliminating the outliers, recalculate the lump size distribution information of the ore to obtain the recalculated lump size distribution information of the ore.

[0060] Specifically, when removing outliers from the lump size distribution information of the ore, first read the particle size characteristic values and their corresponding classifications of each abnormal ore lump one by one from the marked outlier list. Combine the information structure of the lump size distribution information to locate the classification interval where the outlier is located, and remove it from the original classification interval. At the same time, adjust the remaining ore quantity and volume information in the classification interval to ensure the consistency and integrity of the information after outlier removal. The removal process is repeated for all outliers until there are no more outliers in the lump size distribution information. Based on the ore size information after outlier removal, re - statistic and calculate the lump size distribution information of the ore. When recalculating, first count the number of ore lumps in each classification interval, accumulate the volumes of all ore lumps after outlier removal to obtain the volume information of each classification interval. Subsequently, calculate the volume proportion of each classification interval by normalizing the volume information in the classification interval with the total volume of the ore after outlier removal, and recalculate the volume ratio of each classification interval. Organize the recalculated volume ratio information and quantity information and use them as the recalculated ore lump size distribution information. For example, assume that a certain classification interval (such as the medium - sized lump interval of 50 - 200 mm) contains 10 ore lumps with a total volume of 200 cubic meters before outlier removal, and the volume of one abnormal ore lump is 50 cubic meters. After removal, the number of ore lumps in the medium - sized lump interval is adjusted to 9, and the total volume is adjusted to 150 cubic meters. Subsequently, by calculating the volume of the adjusted medium - sized lump interval and the total ore volume (assuming the total ore volume after outlier removal is 1000 cubic meters), the volume proportion of the medium - sized lump interval is recalculated to be 15%. After completion of removal and recalculation, organize the results into the complete recalculated ore lump size distribution information S503: Statistically analyze the recalculated ore lump size distribution information, re - divide the ore proportion of different lump size ranges according to the corrected ore size range, and generate the optimized ore lump size distribution information.

[0061] Specifically, when statistically analyzing the recalculated ore lump size distribution information, first group the recalculated information according to the corrected ore size range. The ore size range is divided according to a preset particle size interval. For example, set small lumps: particle size less than 50 mm, medium lumps: particle size between 50 - 200 mm, and large lumps: particle size greater than 200 mm. Compare the particle size information of each ore lump one by one with the corrected size range and assign it to the corresponding classification interval. After completing the reclassification of the ore lumps, count the number and volume information of the ore in each classification interval. The number of ore is completed by counting each ore lump in the classification interval one by one, and the volume information is obtained by accumulating the volume values of all ore lumps in each classification interval. At the same time, count the volume proportion of each classification interval. The calculation of the volume proportion is completed by normalizing the volume value of the classification interval with the total ore volume after removing outliers, ensuring that the statistical results can accurately reflect the proportion distribution of the ore in each classification interval. Finally, organize the statistical results into optimized ore lump size distribution information. The optimized lump size distribution information includes the number of ore, volume value, and volume proportion in each classification interval, etc. For example, the recalculated information contains 50 ore lumps, the total corrected ore volume is 1000 cubic meters. After statistics, it is found that the number of small lumps (particle size less than 50 mm) is 20, with a total volume of 200 cubic meters, the number of medium lumps (particle size between 50 - 200 mm) is 25, with a total volume of 500 cubic meters, and the number of large lumps (particle size greater than 200 mm) is 5, with a total volume of 300 cubic meters. Calculate the volume proportion according to the ratio of the volume of the classification interval to the total ore volume. Among them, the small lumps account for 20%, the medium lumps account for 50%, and the large lumps account for 30%.

[0062] In one embodiment, as Figure 7 shown, in step S60, that is, according to the optimized lump size distribution information, calculate the large lump rate distribution information of the ore, including: S601: Obtain the volume proportion information of each particle size range of the ore from the optimized lump size distribution information.

[0063] Specifically, when obtaining the volume proportion information of each particle size range of the ore from the optimized lump size distribution information, first extract the volume information of each particle size range from the lump size distribution information, and calculate the volume proportion of each particle size range in combination with the total ore volume. The calculation process is carried out for each particle size interval one by one, specifically including reading the particle size range (such as below 50 mm, 50 - 200 mm, above 200 mm) and the corresponding volume information of each interval, and dividing the volume value of each interval by the total ore volume according to the volume proportion calculation formula to generate the volume proportion information of each particle size range.

[0064] S602: Obtain the target particle size threshold, compare the volume proportion information of each particle size range of the ore with the target particle size threshold, and generate the volume proportion information of the ore particle size greater than the target particle size threshold.

[0065] Specifically, when obtaining the target particle size threshold, first set the target particle size threshold according to process requirements or user input. For example, set the threshold to 200 mm. Subsequently, compare the volume proportion information of each particle size range of the ore with this threshold one by one. For the interval whose particle size range is completely greater than the target particle size threshold, directly classify its volume proportion information as the part greater than the target particle size; for the particle size interval that straddles the target particle size threshold, calculate the volume proportion of the part above the target threshold. This calculation is achieved by analyzing the particle size distribution within this interval, splitting the volume information of the part greater than the target particle size within the interval proportionally and adding it to the classification of the part greater than the target particle size. Finally, integrate all the volume proportion information that meets the conditions to generate the volume proportion information of the ore particle size containing the part greater than the target particle size threshold.

[0066] S603: Normalize the volume proportion information of the ore particle size, calculate the proportion of the volume of the ore greater than the target particle size threshold in the total ore volume, and obtain the bulk rate distribution information of the ore.

[0067] Specifically, when normalizing the volume proportion information of the ore particle size, first accumulate all the volume proportion information greater than the target particle size threshold to obtain the total volume proportion value of the part greater than the target particle size. Subsequently, perform normalization processing on this total volume proportion value and the total volume of the ore to calculate the specific proportion of the bulk rate. During the normalization process, map the volume proportion value to the total ore volume through a normalization formula to ensure the accuracy of the bulk rate calculation. Finally, organize the calculated bulk rate into structured information, and at the same time mark the particle size range and bulk rate composition information of each classification interval to generate the ore bulk rate distribution information containing classification details and the overall bulk rate proportion, providing direct input for subsequent processing process analysis and optimization.

[0068] For example, assume that the volume proportion of the ore with a particle size greater than 200 mm in the particle size distribution is 25%, and the total ore volume is 1000 cubic meters. Then, through cumulative calculation, the volume of the ore with a particle size greater than 200 mm is 250 cubic meters. After normalization, the bulk rate is 250 / 1000 = 25%, and it is organized and output as the bulk rate distribution information.

[0069] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0070] In one embodiment, an ore lump size distribution analysis device is provided, which corresponds one-to-one to the ore lump size distribution analysis method in the above embodiment. As Figure 8 shown, the ore lump size distribution analysis device includes an ore image and depth information acquisition module, an image preprocessing module, a three-dimensional space coordinate generation module, a lump size distribution generation module, a lump size distribution correction module, and a large lump rate calculation and visualization module. The detailed description of each functional module is as follows: The ore image and depth information acquisition module is used to acquire the image information in the ore field and the depth information collected by the sensing device; the image preprocessing module is used to remove the noise of the image information through Gaussian filtering to obtain the image information after noise removal, perform binarization processing on the image information after noise removal to obtain binarized image information, and further perform morphological erosion and dilation operations on the binarized image information to obtain the preprocessed image information; The three-dimensional space coordinate generation module is used to combine the depth information and the preprocessed image information to generate the three-dimensional space coordinate information of the ore through a three-dimensional reconstruction algorithm; The lump size distribution generation module is used to input the three-dimensional space coordinate information of the ore into an intelligent lump size prediction model to generate the lump size distribution information of the ore; The lump size distribution correction module is used to correct the lump size distribution information of the ore by using the standard of the lump size, and eliminate the outliers in the lump size distribution information of the ore to obtain the optimized lump size distribution information; The large lump rate calculation and visualization module is used to calculate the large lump rate distribution information of the ore according to the optimized lump size distribution information, and at the same time display the optimized lump size distribution information and the large lump rate distribution information of the ore through a visualization chart.

[0071] Optionally, the three-dimensional space coordinate generation module includes: An edge extraction and two-dimensional coordinate generation sub-module, which is used to extract the edge pixel coordinate information and the boundary contour information from the preprocessed image information, and generate the initial two-dimensional pixel coordinates according to the edge pixel coordinate information and the boundary contour information; An internal parameter correction and two-dimensional coordinate optimization sub-module, which is used to correct the initial two-dimensional pixel coordinates by using the internal parameter matrix in the three-dimensional reconstruction algorithm to generate the corrected two-dimensional pixel coordinates; An external parameter mapping and three-dimensional space generation sub-module, which is used to combine the corrected two-dimensional pixel coordinates with the depth information by using the external parameter matrix in the three-dimensional reconstruction algorithm to generate the preliminary three-dimensional space coordinate information; A three-dimensional coordinate optimization and outlier elimination sub-module, which is used to detect outliers in the preliminary three-dimensional space coordinate information, eliminate the noise points and outliers, and generate the three-dimensional space coordinate information of the ore after completion.

[0072] Optionally, the internal parameter calibration and two-dimensional coordinate optimization sub-module includes: A two-dimensional pixel optical distortion correction unit, which is used to perform optical distortion correction on two-dimensional pixel coordinates based on the parameters in the internal parameter matrix, eliminate the influence of lens distortion through the following formula, and generate corrected two-dimensional pixel coordinates: where u is the value of the two-dimensional pixel coordinate in the horizontal direction, v is the value of the two-dimensional pixel coordinate in the vertical direction, x′ is the value of the corrected two-dimensional pixel coordinate in the horizontal direction, y′ is the value of the corrected two-dimensional pixel coordinate in the vertical direction, f x is the horizontal focal length parameter of the internal parameter matrix, f y is the vertical focal length parameter of the internal parameter matrix, c x is the horizontal coordinate of the camera principal point, c y is the vertical coordinate of the camera principal point.

[0073] Optionally, the external parameter mapping and three-dimensional space generation sub-module includes: A three-dimensional space coordinate generation unit, which is used to combine the corrected two-dimensional pixel coordinates with depth information based on the rotation matrix and translation vector in the external parameter matrix, and calculate the preliminary three-dimensional space coordinate information standard through the following formula: where X is the value of the three-dimensional space coordinate in the horizontal direction, Y is the value of the three-dimensional space coordinate in the vertical direction, Z is the depth information, R is the rotation matrix in the external parameter matrix, T is the translation vector in the external parameter matrix, and x′ and y′ are the horizontal direction value and vertical direction value of the corrected two-dimensional pixel coordinates respectively.

[0074] Optionally, the lump size distribution generation module includes: A three-dimensional feature extraction sub-module, which is used to extract features from the three-dimensional space coordinate information of the ore and extract ore feature information; an ore classification and size range determination sub-module, which is used to classify the ore according to the ore feature information using a preset ore classification standard, determine the size range of each ore, and generate an ore classification result; A lump size statistics and distribution analysis sub-module, which is used to perform statistical analysis on the ore size ranges in the ore classification result, calculate the quantity ratio of the ore according to different size ranges, and generate lump size distribution information of the ore.

[0075] Optionally, the lump size distribution correction module includes: An outlier identification and elimination sub-module, which is used to compare each ore size in the lump size distribution information of the ore according to a preset lump size standard, and identify outliers that exceed the range of the lump size standard during the comparison process; The lump size distribution information recalculation sub-module is used to remove outliers from the lump size distribution information of the ore to obtain the ore size information after removing outliers, and recalculate the lump size distribution information of the ore based on the ore size information after removing outliers to obtain the recalculated ore lump size distribution information; The lump size statistics and optimization sub-module is used to statistically analyze the recalculated ore lump size distribution information, re-divide the proportion of ores in different lump size ranges according to the corrected ore size range, and generate the optimized ore lump size distribution information.

[0076] Optionally, the large lump rate calculation and visualization module includes: The volume proportion extraction sub-module for particle size ranges is used to obtain the volume proportion information of each particle size range of the ore from the optimized lump size distribution information; The target particle size comparison and volume screening sub-module is used to obtain the target particle size threshold, compare the volume proportion information of each particle size range of the ore with the target particle size threshold, and generate the volume proportion information of the ore particle size greater than the target particle size threshold; The large lump rate calculation and normalization sub-module is used to normalize the volume proportion information of the ore particle size, calculate the proportion of the volume of the ore greater than the target particle size threshold in the total ore volume, and obtain the large lump rate distribution information of the ore.

[0077] For the specific limitations of the ore lump size distribution analysis device, reference can be made to the limitations of the ore lump size distribution analysis method in the above text, which will not be elaborated here. Each module in the above ore lump size distribution analysis device can be implemented in whole or in part through software, hardware and their combination. The above modules can be embedded in the processor of the device in the form of hardware or be independent of it, or can be stored in the memory of the device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0078] In one embodiment, a device is provided. The device can be a server, and its internal structure diagram can be as Figure 9 shown. The device includes a processor, a memory, a network interface and an information database connected through a system bus. Among them, the processor of the device is used to provide computing and control capabilities. The memory of the device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and an information database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements an ore lump size distribution analysis method.

[0079] In one embodiment, a device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: Obtain the image information in the ore yard and the depth information collected by the sensing device; Remove the noise of the image information through Gaussian filtering to obtain the noise-removed image information, perform binarization processing on the noise-removed image information to obtain the binarized image information, and further perform morphological erosion and dilation operations on the binarized image information to obtain the preprocessed image information; Combine the depth information and the preprocessed image information, and generate the three-dimensional space coordinate information of the ore through a three-dimensional reconstruction algorithm; input the three-dimensional space coordinate information of the ore into the intelligent lump size prediction model to generate the lump size distribution information of the ore; Correct the lump size distribution information of the ore using the standard of lump size, and eliminate the outliers in the lump size distribution information of the ore to obtain the optimized lump size distribution information; According to the optimized lump size distribution information, calculate the large lump rate distribution information of the ore, and at the same time display the optimized lump size distribution information and the large lump rate distribution information of the ore through a visualization chart.

[0080] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: Remove the noise of the image information through Gaussian filtering to obtain the noise-removed image information, perform binarization processing on the noise-removed image information to obtain the binarized image information, and further perform morphological erosion and dilation operations on the binarized image information to obtain the preprocessed image information; Combine the depth information and the preprocessed image information, and generate the three-dimensional space coordinate information of the ore through a three-dimensional reconstruction algorithm; input the three-dimensional space coordinate information of the ore into the intelligent lump size prediction model to generate the lump size distribution information of the ore; Correct the lump size distribution information of the ore using the standard of lump size, and eliminate the outliers in the lump size distribution information of the ore to obtain the optimized lump size distribution information; According to the optimized lump size distribution information, calculate the large lump rate distribution information of the ore, and at the same time display the optimized lump size distribution information and the large lump rate distribution information of the ore through a visualization chart.

[0081] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0082] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

[0083] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for analyzing ore fragmentation distribution, characterized in that: The ore size distribution analysis method comprises: Obtain image information within the ore field and depth information collected by sensor equipment; removing noise from the image information by Gaussian filtering to obtain image information after noise removal, binarizing the image information after noise removal to obtain binarized image information, and further performing morphological corrosion and expansion operations on the binarized image information to obtain preprocessed image information; Combining the depth information and the preprocessed image information, generating three-dimensional spatial coordinate information of the ore through a three-dimensional reconstruction algorithm; inputting the three-dimensional spatial coordinate information of the ore into an intelligent particle size prediction model to generate particle size distribution information of the ore; Correcting the particle size distribution information of the ore using the particle size standard, removing abnormal values ​​in the particle size distribution information of the ore, and obtaining optimized particle size distribution information; According to the optimized block size distribution information, the large block rate distribution information of the ore is calculated, and the optimized block size distribution information and the large block rate distribution information of the ore are displayed through a visual chart.

2. The ore size distribution analysis method according to claim 1, characterized in that: The combining of the depth information and the pre-processed image information to generate the three-dimensional spatial coordinate information of the ore by a three-dimensional reconstruction algorithm includes: Extracting edge pixel coordinate information and boundary contour information from the preprocessed image information, and generating initial two-dimensional pixel coordinates according to the edge pixel coordinate information and the boundary contour information; Correcting the initial two-dimensional pixel coordinates using an internal parameter matrix in the three-dimensional reconstruction algorithm to generate corrected two-dimensional pixel coordinates; Combining the corrected two-dimensional pixel coordinates with the depth information using an external parameter matrix in the three-dimensional reconstruction algorithm to generate preliminary three-dimensional space coordinate information; The preliminary three-dimensional spatial coordinate information is subjected to outlier detection to remove noise points and outliers, and upon completion, the three-dimensional spatial coordinate information of the ore is generated.

3. The ore size distribution analysis method according to claim 2, characterized in that: The method of correcting the initial two-dimensional pixel coordinates by using the internal parameter matrix in the three-dimensional reconstruction algorithm to generate corrected two-dimensional pixel coordinates includes: The optical distortion of the two-dimensional pixel coordinates is corrected based on the parameters in the intrinsic parameter matrix, and the influence of lens distortion is eliminated by the following formula to generate the corrected two-dimensional pixel coordinates: Wherein, u is the value of the two-dimensional pixel coordinate in the horizontal direction, v is the value of the two-dimensional pixel coordinate in the vertical direction, x′ is the value of the corrected two-dimensional pixel coordinate in the horizontal direction, y′ is the value of the corrected two-dimensional pixel coordinate in the vertical direction, and f x is the horizontal focal length parameter of the intrinsic parameter matrix, and the f y is the vertical focal length parameter of the intrinsic parameter matrix, and the c x is the horizontal coordinate of the camera principal point, the c y is the longitudinal coordinate of the camera's principal point.

4. The ore size distribution analysis method according to claim 2, characterized in that: The step of combining the corrected two-dimensional pixel coordinates with the depth information using the external parameter matrix in the three-dimensional reconstruction algorithm to generate preliminary three-dimensional space coordinate information includes: Based on the rotation matrix and translation vector in the extrinsic matrix, the corrected two-dimensional pixel coordinates are combined with the depth information, and the preliminary three-dimensional space coordinate information mark is calculated by the following formula: Among them, X is the value of the three-dimensional space coordinate in the horizontal direction, Y is the value of the three-dimensional space coordinate in the vertical direction, Z is the depth information, R is the rotation matrix in the external parameter matrix, T is the translation vector in the external parameter matrix, and x′ and y′ are the horizontal direction value and vertical direction value of the corrected two-dimensional pixel coordinate, respectively.

5. The ore size distribution analysis method according to claim 1, characterized in that: The step of inputting the three-dimensional spatial coordinate information of the ore into the intelligent particle size prediction model to generate the particle size distribution information of the ore includes: Performing feature extraction on the three-dimensional spatial coordinate information of the ore to extract ore feature information; According to the ore characteristic information, the ore is classified using a preset ore classification standard, the size range of each ore is determined, and an ore classification result is generated; A statistical analysis is performed on the ore size range in the ore classification result, and the quantity ratio of the ore is calculated according to different size ranges to generate the block size distribution information of the ore.

6. The ore size distribution analysis method according to claim 1, characterized in that: The method of correcting the particle size distribution information of the ore by using the particle size standard, removing abnormal values ​​in the particle size distribution information of the ore, and obtaining optimized particle size distribution information includes: According to the preset block size standard, the ore sizes in the block size distribution information of the ore are compared one by one, and in the comparison process, the abnormal values ​​exceeding the block size standard range are identified; The abnormal value is removed from the particle size distribution information of the ore to obtain the ore size information after the abnormal value is removed, and the particle size distribution information of the ore is recalculated according to the ore size information after the abnormal value is removed to obtain the recalculated ore particle size distribution information; The recalculated ore fragmentation distribution information is statistically analyzed, and the proportion of ores in different fragmentation ranges is re-divided according to the corrected ore size range to generate optimized ore fragmentation distribution information.

7. The ore size distribution analysis method according to claim 1, characterized in that: The step of calculating the large block rate distribution information of the ore according to the optimized block size distribution information includes: Obtaining volume proportion information of each particle size range of the ore from the optimized block size distribution information; Obtaining a target particle size threshold, comparing the volume percentage information of each particle size range of the ore with the target particle size threshold, and generating volume percentage information of ore particle sizes greater than the target particle size threshold; The ore particle size volume ratio information is normalized and sorted, and the volume ratio of the ore larger than the target particle size threshold in the total ore volume is calculated to obtain the bulk rate distribution information of the ore.

8. An ore size distribution analysis device, characterized in that: The ore size distribution analysis device comprises: An ore image and depth information acquisition module is used to acquire image information in the ore field and depth information collected by the sensor equipment; an image preprocessing module is used to remove noise from the image information by Gaussian filtering to obtain image information after noise removal, perform binarization processing on the image information after noise removal to obtain binarized image information, and further perform morphological corrosion and expansion operations on the binarized image information to obtain preprocessed image information; A three-dimensional space coordinate generation module, used to generate three-dimensional space coordinate information of the ore by combining the depth information and the pre-processed image information through a three-dimensional reconstruction algorithm; A fragmentation distribution generation module, used for inputting the three-dimensional spatial coordinate information of the ore into the intelligent fragmentation prediction model to generate fragmentation distribution information of the ore; A particle size distribution correction module is used to correct the particle size distribution information of the ore using a particle size standard, remove abnormal values ​​in the particle size distribution information of the ore, and obtain optimized particle size distribution information; The large block rate calculation and visualization module is used to calculate the large block rate distribution information of the ore according to the optimized block size distribution information, and to display the optimized block size distribution information and the large block rate distribution information of the ore through a visualization chart.

9. A device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the ore fragmentation distribution analysis method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the ore fragmentation distribution analysis method according to any one of claims 1 to 7 are implemented.

Images