Step line detection method and device for strip mine and storage medium

Abstract

The invention discloses a step line detection method and device for a strip mine and a storage medium. The step line detection method comprises the steps of obtaining point cloud data of the strip mine; extracting target point cloud data in the point cloud data, wherein the target point cloud data comprises ground points in the point cloud data; step line feature points meeting the constraint condition of each dimension in the ground points are selected based on multiple constraint conditions of different dimensions; determining a feature point type of each step line feature point based on the satisfied constraint condition, wherein the feature point type comprises a strong feature point and a weak feature point; and taking the strong feature points as control points, taking the weak feature points as reference points, and performing curve fitting on the strong feature points and the weak feature points to obtain step line information of the strip mine. According to the scheme, the screening accuracy of the step line feature points can be improved, and accurate detection of the step line of the strip mine can be realized through a method of performing curve fitting by combining the strong feature points and the weak feature points.

Description

Technical Field

[0001] This application relates to the field of slope data processing technology, specifically to a method, apparatus, and storage medium for detecting bench lines in open-pit mines. Background Technology

[0002] In open-pit mining, the bench line, as a core geographical element characterizing the spatial morphology of the mining bench, requires precise acquisition and dynamic updating. This is crucial for optimizing blasting parameters, intelligently scheduling transportation routes, and providing early warnings of slope risks. Currently, the widely used RTK discrete point measurement technology suffers from insufficient data acquisition density and incomplete 3D feature representation, making it difficult to accurately depict the continuous spatial morphology of complex mining benches. Meanwhile, manual plotting methods based on 2D DEMs are limited by planar interpretation modes, easily leading to misjudgments of terrain features during 3D spatial relationship reconstruction, resulting in significant deviations between model accuracy and actual mining conditions. Existing technologies generally face common technical bottlenecks such as high levels of human intervention, weak 3D feature analysis capabilities, and poor data update timeliness, severely hindering the intelligentization of mine production management. Summary of the Invention

[0003] The purpose of this application is to provide a method, apparatus, and storage medium for bench line detection in open-pit mines, in order to solve the problems of high reliance on manual labor and limited efficiency and accuracy in bench line detection in the prior art.

[0004] To achieve the above objectives, the first aspect of this application provides a method for detecting bench lines in open-pit mines, the method comprising: Obtain point cloud data from open-pit mines; Extract the target point cloud data from the point cloud data, which includes ground points in the point cloud data; Based on constraints of multiple different dimensions, step line feature points in the ground points that satisfy the constraints of each dimension are selected. The feature point type of each step line feature point is determined based on the satisfied constraints. The feature point types include strong feature points and weak feature points. Strong feature points are used as control points, and weak feature points are used as reference points. Curve fitting is performed on the strong and weak feature points to obtain the bench line information of the open-pit mine.

[0005] In embodiments of this application, the method further includes: performing layered processing on the target point cloud data to generate multiple layered intervals; determining the overlapping area between any two adjacent layered intervals; determining target ground points with different feature point types among the ground points included in each overlapping area; and correcting the feature point type of the target ground point according to the feature point types of all ground points in the neighborhood of the target ground point.

[0006] In the embodiments of this application, the layering process of the target point cloud data to generate multiple layered intervals includes: obtaining the height range of the target point cloud data in the elevation data field; determining the layer thickness of each layered interval based on the height range and the preset layer number; and generating multiple layered intervals of the target point cloud data using a sliding window based on the layer thickness and the preset overlap rate.

[0007] In embodiments of this application, the method further includes: after generating multiple layered intervals, determining the slope change data of each layered interval; adjusting the layer thickness of each layered interval according to the slope change data; or for each layered interval, if the layer thickness of the layered interval is less than a preset layer thickness threshold, merging the layered interval into a layered interval adjacent to the layered interval, or allowing the layered interval to be less than the preset layer thickness threshold.

[0008] In the embodiments of this application, correcting the feature point type of the target ground point based on the feature point types of all ground points in the neighborhood of the target ground point includes: determining the feature point type of each ground point in the neighborhood of the target ground point; determining the proportion of ground points of each feature point type in the neighborhood of the target ground point; and correcting the feature point type of the target ground point to the feature point type corresponding to the maximum proportion.

[0009] In the embodiments of this application, determining the feature point type of each step line feature point based on the satisfied constraints includes: determining the feature point type of step line feature points that satisfy constraints in all dimensions as strong feature points; and determining the feature point type of step line feature points that satisfy constraints in at least one dimension but not all dimensions as weak feature points.

[0010] In the embodiments of this application, determining the feature point type of each step line feature point based on the satisfied constraints includes: for each step line feature point, determining the weight value corresponding to the constraint condition of each dimension; for each step line feature point, performing a weighted sum based on all weight values ​​and the constraint conditions satisfied by the step line feature point to determine the feature point type of the step line feature point.

[0011] In the embodiments of this application, selecting step line feature points among ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions includes: for each first ground point, determining the local normal vector of each second ground point in the neighborhood of the first ground point; determining the angle between the normal vector of each local normal vector and a preset direction, where the preset direction is the direction perpendicular to the ground; for each first ground point, determining whether there is a sudden change in the local normal vector of the first ground point in the neighborhood of all second ground points corresponding to the first ground point; for any first ground point, if the angle between the normal vectors corresponding to the first ground point is less than a preset angle threshold and there is a sudden change in the local normal vector of the first ground point, determining the first ground point as a step line feature point that satisfies the constraint of the normal vector angle.

[0012] In the embodiments of this application, selecting step line feature points among ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions includes: performing surface fitting on all second ground points in the neighborhood of each first ground point using the moving least squares method to generate a fitted surface; determining the Gaussian curvature and average curvature of each first ground point on the fitted surface corresponding to that ground point; fusing the Gaussian curvature and average curvature corresponding to each first ground point to obtain the composite curvature of each first ground point; and for each first ground point, if the composite curvature of the first ground point is less than a preset curvature threshold, determining the first ground point as a step line feature point that satisfies the curvature constraint.

[0013] In the embodiments of this application, selecting step line feature points among ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions includes: performing gridding processing on the target point cloud data and filling each grid cell with an elevation value; for each ground point, determining the elevation gradient of the ground point based on the elevation values ​​of the grid cells included in the neighborhood of the ground point using an edge detection algorithm; for each ground point, determining an elevation gradient threshold set for the ground point based on the elevation values ​​of the grid cells included in the neighborhood of the ground point; and for each ground point, determining the ground point as a step line feature point that satisfies the elevation gradient constraint if the elevation gradient of the ground point is greater than its corresponding elevation gradient threshold.

[0014] In embodiments of this application, extracting target point cloud data from point cloud data includes: performing progressive morphological filtering on the acquired open-pit mine point cloud data to extract target point cloud data; for each first ground point in the target point cloud data, determining the normal vector angle and distance between the first ground point and each second ground point in the neighborhood of the first ground point; for each first ground point, determining corresponding first weights and second weights based on the normal vector angle and distance between the first ground point and each second ground point; for each first ground point, performing bilateral filtering on the second ground points in the neighborhood of the first ground point based on all first weights and all second weights corresponding to the first ground point to correct the target point cloud data.

[0015] A second aspect of this application provides a bench line detection device for open-pit mines, comprising: The memory is configured to store instructions; The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement the aforementioned method for detecting step lines in open-pit mines.

[0016] A third aspect of this application provides a machine-readable storage medium storing instructions for causing a machine to perform the above-described method for detecting bench lines in open-pit mines.

[0017] This technical solution integrates multi-dimensional terrain feature analysis with automated extraction algorithms, overcoming the limitations of traditional manual measurement methods. By using multi-feature collaborative analysis to distinguish feature points along the bench lines, it further classifies the feature point types of ground points in point cloud data, improving the accuracy of data filtering. Through curve fitting combining strong and weak feature points, accurate detection of bench lines in open-pit mines can be achieved. This technical solution provides an innovative approach to spatial data acquisition, 3D model construction, and mining decision support during open-pit mining, and has significant engineering application value in promoting the evolution of mine surveying technology towards intelligence and real-time capabilities.

[0018] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 The illustration shows a flowchart of a method for detecting bench lines in an open-pit mine according to an embodiment of this application. Figure 2This illustration schematically shows a flowchart of a multi-feature fusion analysis process according to an embodiment of this application; Figure 3 This illustration schematically shows a process diagram of a dynamic elevation layering method for cutting point clouds according to an embodiment of this application; Figure 4 This schematic diagram illustrates a structural block diagram of a step line detection device for an open-pit mine according to an embodiment of this application; Figure 5 This schematic diagram illustrates a structural block diagram of a step line detection device for an open-pit mine according to an embodiment of this application; Figure 6 The illustration shows a schematic diagram of the structure of a computer device according to an embodiment of the present application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0021] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0022] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0023] Figure 1 The illustration schematically shows a flowchart of a bench line detection method for open-pit mines according to an embodiment of this application. Figure 1As shown in the figure, this application provides a method for detecting bench lines in open-pit mines, which may include the following steps.

[0024] S102, acquire point cloud data of the open-pit mine.

[0025] Specifically, point cloud data from open-pit mines can be acquired using airborne lidar and ground-based 3D laser scanners, thereby accurately obtaining high-density 3D spatial data of the mine surface and slopes.

[0026] S104, Extract target point cloud data from point cloud data, the target point cloud data includes ground points in point cloud data.

[0027] It's understandable that the bench lines in open-pit mines refer to the horizontal layers of a certain thickness that divide the ore into during mining operations. These layers form a stepped spatial distribution during extraction. Each such stepped structure is called a bench, and the top and bottom lines of the benches constitute the bench lines. Therefore, extracting bench lines first requires filtering out non-ground points from the point cloud data, retaining only the ground points.

[0028] In embodiments of this application, extracting target point cloud data from point cloud data includes: performing progressive morphological filtering on the acquired open-pit mine point cloud data to extract target point cloud data; for each first ground point in the target point cloud data, determining the normal vector angle and distance between the first ground point and each second ground point in the neighborhood of the first ground point; for each first ground point, determining corresponding first weights and second weights based on the normal vector angle and distance between the first ground point and each second ground point; for each first ground point, performing bilateral filtering on the second ground points in the neighborhood of the first ground point based on all first weights and all second weights corresponding to the first ground point to correct the target point cloud data.

[0029] It is understandable that filtering out non-ground points in point cloud data can begin by performing progressive morphological filtering on the point cloud data. Specifically, for each point, based on the elevation data of the point cloud, the terrain slope value of that point is calculated through the elevation differences of multiple points in its neighborhood. Furthermore, a sliding window is used to dynamically acquire the slope values ​​of each region in the point cloud data to analyze the slope distribution characteristics of each region. Further, based on the mapping relationship between slope values ​​and the size of the morphological filtering window, sliding windows of different sizes are set to remove non-ground points such as mining assets, equipment, and vegetation from the point cloud data, extracting the ground points.

[0030] Furthermore, to maintain the integrity of the terrain structure of the target point cloud data, further filtering can be performed. Specifically, for each first ground point in the target point cloud data, the normal vector of the first ground point is calculated, and the normal vector of each second ground point in its neighborhood is also calculated to determine the angle between the normal vectors of each second ground point and the first ground point, and the distance between them can also be determined. It is understood that the terms "first ground point" and "second ground point" are relative. Further, for each first ground point, a corresponding first weight is determined based on the angle between the normal vectors of the first ground point and each second ground point, and a corresponding second weight is determined based on the distance between them. Based on all the first weights and all the second weights, each first ground point can be subjected to bilateral filtering on the second ground points in its neighborhood to correct the target point cloud data. Specifically, points with excessively low weights can be removed, or the positions of points can be adjusted. For example, the normal directions of flat areas are similar, while noise points may exhibit abrupt changes in normal direction, thus being identified as outliers. The smaller the angle between the normal vectors (indicating surface continuity), the greater the weight; conversely, if the angle between the normal vectors is too large (e.g., in noise or at edges), the weight decreases. By processing target point cloud data using bilateral filtering, spatial proximity and the angle between the normal vectors are used as criteria to filter out noise points, thus eliminating outliers while maintaining the integrity of the terrain structure.

[0031] S106, based on multiple constraints of different dimensions, select the step line feature points among the ground points that satisfy the constraints of each dimension.

[0032] It is understood that step line feature points refer to key points that possess the boundary and shape of a step. Step line feature points can be filtered based on constraints. In the embodiments of this application, step line feature points with constraints in different dimensions can be filtered out. Furthermore, step line feature points can be further analyzed through multi-dimensional feature synergy to improve the accuracy of data filtering. Specifically, constraints include, but are not limited to, dimensions such as the angle between normal vectors, curvature, and elevation gradient.

[0033] In the embodiments of this application, selecting step line feature points among ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions includes: for each first ground point, determining the local normal vector of each second ground point in the neighborhood of the first ground point; determining the angle between the normal vector of each local normal vector and a preset direction, where the preset direction is the direction perpendicular to the ground; for each first ground point, determining whether there is a sudden change in the local normal vector of the first ground point in the neighborhood of all second ground points corresponding to the first ground point; for any first ground point, if the angle between the normal vectors corresponding to the first ground point is less than a preset angle threshold and there is a sudden change in the local normal vector of the first ground point, determining the first ground point as a step line feature point that satisfies the constraint of the normal vector angle.

[0034] It's understandable that the calculation of local normal vectors can utilize Principal Component Analysis (PCA). The covariance matrix of the neighborhood point set is calculated, and then eigenvalue decomposition is used to find the eigenvector corresponding to the smallest eigenvalue, i.e., the normal vector. It's important to note that the direction of the normal vectors within the neighborhood needs to be unified, for example, towards the viewpoint or through other methods. It's understandable that for stepped terrain, the elevation at the edge of the step may change abruptly, or the direction of the normal vector may change drastically within the neighborhood. Therefore, the degree of directional change can be quantified by calculating the standard deviation of the local normal vectors within the neighborhood. On the other hand, if the first ground point is a feature point of the step line, then according to the geometric rules of the step line, the normal vector of the first ground point needs to be close to the vertical plane. Therefore, the angle between the normal vector and the vertical direction (e.g., the Z-axis) can be used to determine whether the first ground point is close to the vertical plane. If the angle between the normal vector and the Z-axis is less than a preset angle threshold, then the point is considered to belong to a vertical plane structure, such as the elevation of the step. Combining the two conditions above, points that belong to the vertical plane and exhibit abrupt changes within their neighborhood can be selected. These points may be located at the edges of steps, forming step line features. It is understandable that the standard deviation of the local normal vector at the edge of a real step is usually small (e.g., less than 25 degrees), while the noise region may be larger (e.g., exceeding 40 degrees). Therefore, a preset angle threshold can be set according to specific circumstances to improve the accuracy of step line detection.

[0035] In the embodiments of this application, for each point P in the point cloud, the K nearest neighbor points are found to form a neighborhood. This step typically uses the K-nearest neighbor algorithm or a spatial partitioning method, such as a KD-tree, to quickly find neighboring points, which provides a foundation for subsequent calculations.

[0036] The aforementioned method can accurately identify the location of the bench line, providing an important basis for open-pit mine planning and safety monitoring. Secondly, by combining the direction of the normal vector and the elevation change, the influence of noise points can be effectively reduced, improving the robustness of bench line detection. Utilizing efficient data structures such as the K-nearest neighbor algorithm or KD-trees to find neighboring points can significantly improve computational efficiency, enabling bench line detection to be performed in real-time or near real-time conditions, thus jointly promoting the intelligent and automated development of open-pit mining technology.

[0037] In the embodiments of this application, selecting step line feature points among ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions includes: performing surface fitting on all second ground points in the neighborhood of each first ground point using the moving least squares method to generate a fitted surface; determining the Gaussian curvature and average curvature of each first ground point on the fitted surface corresponding to that ground point; fusing the Gaussian curvature and average curvature corresponding to each first ground point to obtain the composite curvature of each first ground point; and for each first ground point, if the composite curvature of the first ground point is less than a preset curvature threshold, determining the first ground point as a step line feature point that satisfies the curvature constraint.

[0038] Specifically, for any point P in the point cloud, its K nearest neighbors are found through spatial search (such as a KD-tree), forming a local region (K-neighborhood). The local neighborhood provides contextual information about the geometric features surrounding point P, forming the basis for subsequent calculations. Within the K-neighborhood, a smooth surface is fitted using weighted least squares (MLS). MLS introduces a distance-dependent weighting function, making the fitted surface more adaptable to local geometry and robust to noise. Specifically, the weighting function can be a Gaussian kernel. Further, the Gaussian curvature and mean curvature of each first ground point on its fitted surface are calculated. Fusing Gaussian and mean curvature information enhances sensitivity to features. Specifically, this can be achieved through linear combination, incorporating corresponding weighting coefficients to obtain a composite curvature. Alternatively, it can be achieved through geometric averaging or by dynamically adjusting the combination method based on data distribution to obtain a composite curvature. A critical value (preset curvature threshold) is determined experimentally or statistically. When the composite curvature exceeds the preset curvature threshold, the first ground point is determined to be a step line feature point. By using curvature thresholding, features such as step edges can be identified because feature points are usually located in regions of abrupt curvature changes (such as transitions from a plane to a vertical plane).

[0039] In the embodiments of this application, selecting step line feature points among ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions includes: performing gridding processing on the target point cloud data and filling each grid cell with an elevation value; for each ground point, determining the elevation gradient of the ground point based on the elevation values ​​of the grid cells included in the neighborhood of the ground point using an edge detection algorithm; for each ground point, determining an elevation gradient threshold set for the ground point based on the elevation values ​​of the grid cells included in the neighborhood of the ground point; and for each ground point, determining the ground point as a step line feature point that satisfies the elevation gradient constraint if the elevation gradient of the ground point is greater than its corresponding elevation gradient threshold.

[0040] Specifically, the target point cloud data is meshed, converting the disordered point cloud into a regular two-dimensional elevation matrix. The mesh size can be selected based on the point cloud density. For each mesh cell, the maximum, minimum, or average elevation values ​​of its internal points can be used to fill the elevation values. An edge detection algorithm is used to determine the elevation gradient of a ground point based on the elevation values ​​of the mesh cells within its neighborhood. Specifically, the Sobel operator can be used to calculate the elevation gradient. The Sobel operator is commonly used for edge detection in image processing, identifying locations of abrupt changes in pixel values ​​by calculating the gradient. In a point cloud, the elevation gradient may refer to the rate of change along the Z-axis, i.e., the change in terrain height. Adapting the Sobel operator to 3D point cloud data may require projecting the point cloud onto a 2D mesh or processing it directly in 3D space to capture regions of abrupt elevation changes (such as the edges of steps). This improves computational efficiency and is suitable for processing large-scale point cloud data. Alternatively, the elevation values ​​of the mesh cells within the neighborhood of each ground point can be divided using a box plot method. Specifically, boxplots are used to represent the distribution of statistical data, and quartiles and outlier detection are used to determine the elevation gradient threshold for each ground point. Then, for each ground point, if the elevation gradient of the ground point is greater than its corresponding elevation gradient threshold, the step line feature points are obtained under the constraint of elevation gradient detection.

[0041] S108, determine the feature point type of each step line feature point based on the satisfied constraints. The feature point type includes strong feature points and weak feature points.

[0042] It's understandable that the step line feature points selected based on different constraints satisfy at least one constraint. However, some of these feature points may satisfy more than one constraint. Therefore, a step line feature point that satisfies multiple constraints can be called a strong feature point. A step line feature point that satisfies only one constraint can be called a weak feature point. Ground points that do not satisfy any constraints are considered non-feature points. It should be noted that the classification of strong and weak feature points is not necessarily based on the number of dimensions satisfied by the constraints. (Refer to...) Figure 2 Alternatively, the importance of different constraints can be considered, and a weighted fusion model function can be used to determine the type of feature point. Specifically, different weights can be set for different constraints, and then a weighted sum can be performed on the same feature point based on different weights to determine its type. Compared to a single dimension, multi-feature fusion analysis can accurately identify the set of step feature points.

[0043] S110 uses strong feature points as control points and weak feature points as reference points to perform curve fitting on the strong and weak feature points to obtain the bench line information of the open-pit mine.

[0044] Specifically, strong feature points can be used as control points in the step line fitting calculation, while weak feature points can be used as reference points for step line fitting correction. This allows for the fitting of a relatively accurate open-pit mine step line. Specifically, a third-order Bézier curve algorithm can be used for curve fitting. The third-order Bézier curve algorithm has good smoothness and continuity, and can well simulate the actual shape of the open-pit mine step line. During the fitting process, strong feature points, as control points, play a key role in guiding the curve trend, ensuring that the fitted step line is consistent with the actual step line trend. Weak feature points, as reference points, can fine-tune the fitted step line, further improving the fitting accuracy. By combining strong and weak feature points for curve fitting, accurate detection of open-pit mine step lines can be achieved.

[0045] This technical solution integrates multi-dimensional terrain feature analysis with automated extraction algorithms, overcoming the limitations of traditional manual measurement methods. By using multi-feature collaborative analysis to distinguish feature points along the bench lines, it further classifies the feature point types of ground points in point cloud data, improving the accuracy of data filtering. Through curve fitting combining strong and weak feature points, accurate detection of bench lines in open-pit mines can be achieved. This technical solution provides an innovative approach to spatial data acquisition, 3D model construction, and mining decision support during open-pit mining, and has significant engineering application value in promoting the evolution of mine surveying technology towards intelligence and real-time capabilities.

[0046] In the embodiments of this application, after obtaining the aforementioned step line information, a length threshold filtering mechanism can be introduced. Specifically, by calculating the Euclidean distance between the start and end points of the fitted curve, line segments with an Euclidean distance less than the step line length threshold are removed, thereby eliminating invalid step line segments and finally obtaining the optimized open-pit mine cloud step line information.

[0047] In embodiments of this application, the method further includes: performing layered processing on the target point cloud data to generate multiple layered intervals; determining the overlapping area between any two adjacent layered intervals; determining target ground points with different feature point types among the ground points included in each overlapping area; and correcting the feature point type of the target ground point according to the feature point types of all ground points in the neighborhood of the target ground point.

[0048] It's understandable that the target point cloud data, obtained after filtering, can be dynamically layered into multiple layered intervals based on elevation. After layering, each ground point is processed using the ground points within its own layered interval as the data basis, meaning the data processing between layered intervals is independent. Layering can require a certain overlap between layers to prevent broken or misaligned step lines due to improper layering. It's understandable that dividing the target point cloud data into multiple layered intervals allows for parallel data processing, improving efficiency. However, this approach can lead to fragmentation of point cloud data in the boundary regions between layered intervals. For example, for ground point 'a' in layered interval 1, if we want to calculate the step line feature points of ground point 'a', even if points in layered interval 2 are also within its neighborhood, those points cannot participate in the calculation. Therefore, in overlapping areas between adjacent layered intervals, the same ground point might generate different feature point types in different intervals. Therefore, these ground points can be used as target ground points for filtering. Furthermore, the feature point type of the target ground point can be refined based on the feature point types of all ground points in its neighborhood. This improves both data processing efficiency and robustness.

[0049] In the embodiments of this application, correcting the feature point type of the target ground point based on the feature point types of all ground points in the neighborhood of the target ground point includes: determining the feature point type of each ground point in the neighborhood of the target ground point; determining the proportion of ground points of each feature point type in the neighborhood of the target ground point; and correcting the feature point type of the target ground point to the feature point type corresponding to the maximum proportion.

[0050] Specifically, after identifying target ground points with different feature point types within the overlapping region, the neighborhood of each target ground point is determined. It's understood that the selected neighborhood here is not affected by the layered processing but is based on the overall target point cloud data. Therefore, based on the proportion of ground points of different feature point types within the neighborhood of the target ground point, the feature point type corresponding to the largest proportion is determined as the feature point type of the target ground point.

[0051] In another implementation, a neighborhood of the target ground point is selected. Based on the spatial distance between the target ground point and other ground points within the neighborhood, a weight is assigned to each ground point. It's understood that ground points with greater spatial distance receive lower weights. Then, combining this with the feature point type of each ground point, a comprehensive weight value for each feature point type within the neighborhood is calculated. Specifically, this value can be obtained by directly summing the weights. Finally, the feature point type with the highest comprehensive weight value is selected as the feature point type of the target ground point.

[0052] Another implementation scheme addresses the issue of inconsistent feature values ​​for step line feature points within overlapping layers due to differences in thresholds between upper and lower layers. This can be achieved through dynamic correction using an inverse distance weighted interpolation algorithm. Specifically, firstly, within each upper and lower layer, representative values ​​are calculated using inverse distance weighted interpolation, centered on the target ground point and incorporating its own and its neighboring feature points. Then, a layered reliability index is calculated based on the dispersion of the neighboring feature values ​​in each layer. Finally, the interpolation results from the upper and lower layers are weighted and fused according to the reliability ratio.

[0053] The above-mentioned methods can all correct the feature point type of the target ground point, making full use of information from neighboring ground points to more accurately determine the feature point type of the target ground point. This not only improves the accuracy of bench line detection but also provides more reliable data support for subsequent mining operations. At the same time, these methods also have a certain degree of flexibility and applicability, and can be appropriately adjusted and optimized according to specific open-pit mine environments and needs.

[0054] In the embodiments of this application, the layering process of the target point cloud data to generate multiple layered intervals includes: obtaining the height range of the target point cloud data in the elevation data field; determining the layer thickness of each layered interval based on the height range and the preset layer number; and generating multiple layered intervals of the target point cloud data using a sliding window based on the layer thickness and the preset overlap rate.

[0055] refer to Figure 3, the total height range is determined according to the minimum and maximum values of the entire target point cloud data in the elevation quantity field. Further, the layer thickness of the stratification interval can be determined by combining the preset number of layers with the height range. Further, based on the required overlap rate, the sliding window method is used to generate the stratification interval, and the length supplemented by each sliding is the layer thickness minus the overlap degree. It can be understood that the overlap rate can ensure that information near the boundary will not be lost due to strict division during segmentation, especially in places with large terrain changes, and the overlap can help capture more details. The setting of the overlap rate needs to be adjusted according to the actual situation of the open-pit mine to ensure that the bench lines between each layer can be effectively connected while avoiding excessive data redundancy and improving data processing efficiency. In addition, the stratification process also needs to consider the actual height, slope of the bench and the complexity of the terrain to achieve more accurate bench line detection.

[0056] In an embodiment of the present application, the method further includes: after generating multiple stratification intervals, determining the slope change data of each stratification interval; adjusting the layer thickness of each stratification interval according to the slope change data; or for each stratification interval, when the layer thickness of the stratification interval is less than the preset layer thickness threshold, merging the stratification interval into the stratification interval adjacent to the stratification interval, or allowing the stratification interval to be less than the preset layer thickness threshold.

[0057] Reference Figure 3 , the stratification process can also dynamically adjust the thickness of each layer according to the characteristics of the mine terrain. A smaller layer thickness can be used in areas with剧烈 elevation changes, and a larger layer thickness can be used in areas with gentle changes. For the situation where the top or bottom layer is less than a complete window, extension or merging with the adjacent layer can be selected to ensure data integrity and obtain the stratified point cloud data. The above solution establishes an adaptive stratification model based on the elevation distribution of point clouds, and through a dynamic stratification threshold control mechanism, realizes thin layer segmentation in steep slope areas and thick layer division in gentle areas, ensuring a stratification effect that matches the terrain characteristics.

[0058] In a specific embodiment, the terrain-adaptive stratification can be generated by using the sliding window method according to the maximum elevation difference of the entire point cloud elevation data, the preset layer thickness H, and the required overlap rate n. Among them, the window is slid upward from the lowest elevation with an initial step length S = H·x(1 - n). When the elevation change rate within the window exceeds the threshold, the layer thickness is dynamically reduced to kH(0 < k < 1), and in the gentle area, it is enlarged to mH(m > 1).

[0059] The above technical solution employs a progressive morphological filtering algorithm with terrain slope adaptive window adjustment, combined with a bilateral filtering method optimized by dynamic thresholds, to filter out non-ground point cloud features while maintaining the integrity of the terrain structure. A dynamic layering model is constructed based on the elevation distribution characteristics of the point cloud. A layering threshold control mechanism enables thin-layer segmentation in steep slope areas and thick-layer division in gentle areas, adapting to the layering requirements of different terrain features. Based on terrain slope adaptive filtering and dynamic layering control, a progressive morphological filtering algorithm dynamically adjusts the window size, automatically optimizing the layering threshold in conjunction with terrain undulations to solve the problems of over-segmentation in steep slope areas and under-segmentation in gentle areas, improving the layering accuracy and structural integrity under complex terrain. A three-dimensional joint discrimination criterion of normal vector angle constraint, curvature threshold screening, and elevation gradient detection is proposed. Multi-dimensional features are integrated to identify step feature point sets, and multi-feature collaborative analysis distinguishes step line feature points, improving the accuracy of data filtering. A neighborhood mean-weighted fusion algorithm is designed to process feature points in overlapping areas between upper and lower layers. Combined with a third-order Bézier curve parametric modeling method, the spatial distribution and geometric morphology of the step lines are optimized, achieving smooth connection of cross-layer features and correction of jagged edges, thus enhancing the continuity of 3D reconstruction. Furthermore, a length threshold filtering mechanism is used to remove redundant line segments. This solution goes beyond a single data processing technique, comprehensively utilizing multiple methods to form a complete step line extraction solution based on open-pit mine point cloud data. It can meet the needs of users in different scenarios and has high practicality and applicability.

[0060] Figure 1 This is a flowchart illustrating a bench line detection method for open-pit mines in one embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0061] Figure 4 This schematically illustrates a structural block diagram of a bench line detection device for an open-pit mine according to an embodiment of this application. Figure 4 As shown in the figure, this application provides a bench line detection device for open-pit mines, which may include: The memory is configured to store instructions; The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement the aforementioned step line detection method for open-pit mines.

[0062] Specifically, in this embodiment of the application, the processor can be configured as follows: Acquire point cloud data of the open-pit mine; extract target point cloud data from the point cloud data, which includes ground points in the point cloud data; select step line feature points from the ground points that satisfy the constraints of each dimension based on multiple constraints; determine the feature point type of each step line feature point based on the satisfied constraints, including strong feature points and weak feature points; use strong feature points as control points and weak feature points as reference points, and perform curve fitting on the strong and weak feature points to obtain the step line information of the open-pit mine.

[0063] In embodiments of this application, the processor may also be configured to: The target point cloud data is processed into layers to generate multiple layered intervals; the overlapping area between any two adjacent layered intervals is determined; the target ground points with different feature point types among the ground points included in each overlapping area are identified; and the feature point type of the target ground point is corrected based on the feature point types of all ground points in the neighborhood of the target ground point.

[0064] In embodiments of this application, the processor may also be configured to: Obtain the height range of the target point cloud data in the elevation data field; determine the layer thickness of each layer interval based on the height range and the preset number of layers; and generate multiple layer intervals of the target point cloud data using a sliding window based on the layer thickness and the preset overlap rate.

[0065] In embodiments of this application, the processor may also be configured to: After generating multiple layered intervals, determine the slope change data for each layered interval; adjust the layer thickness of each layered interval based on the slope change data; or for each layered interval, if the layer thickness of the layered interval is less than a preset layer thickness threshold, merge the layered interval into the layered interval adjacent to the layered interval, or allow the layered interval to be less than the preset layer thickness threshold.

[0066] In embodiments of this application, the processor may also be configured to: Determine the feature point type of each ground point in the neighborhood of the target ground point; determine the proportion of ground points of each feature point type in the neighborhood of the target ground point; and correct the feature point type of the target ground point to the feature point type corresponding to the maximum proportion.

[0067] In the embodiments of this application, determining the feature point type of each step line feature point based on the satisfied constraints includes: determining the feature point type of step line feature points that satisfy constraints in all dimensions as strong feature points; and determining the feature point type of step line feature points that satisfy constraints in at least one dimension but not all dimensions as weak feature points.

[0068] In embodiments of this application, the processor may also be configured to: For each step line feature point, determine the weight value corresponding to the constraint condition of each dimension; for each step line feature point, perform a weighted sum based on all weight values ​​and the constraint conditions satisfied by the step line feature point to determine the feature point type of the step line feature point.

[0069] In embodiments of this application, the processor may also be configured to: For each first ground point, determine the local normal vector of each second ground point in the neighborhood of the first ground point; determine the angle between each local normal vector and the preset direction, where the preset direction is perpendicular to the ground; for each first ground point, determine whether there is a sudden change in the local normal vector of the first ground point in the neighborhood of all the second ground points corresponding to the first ground point; for any first ground point, if the angle between the normal vectors corresponding to the first ground point is less than the preset angle threshold and there is a sudden change in the local normal vector of the first ground point, determine the first ground point as a step line feature point that satisfies the normal vector angle constraint condition.

[0070] In embodiments of this application, the processor may also be configured to: The moving least squares method is used to fit a surface to all second ground points in the neighborhood of each first ground point to generate a fitted surface; the Gaussian curvature and mean curvature of each first ground point on the fitted surface are determined; the Gaussian curvature and mean curvature of each first ground point are fused to obtain the composite curvature of each first ground point; for each first ground point, if the composite curvature of the first ground point is less than a preset curvature threshold, the first ground point is determined to be a step line feature point that satisfies the curvature constraint condition.

[0071] In embodiments of this application, the processor may also be configured to: The target point cloud data is processed into a grid, and each grid cell is filled with an elevation value. For each ground point, the elevation gradient of the ground point is determined based on the elevation values ​​of the grid cells in its neighborhood, using an edge detection algorithm. For each ground point, an elevation gradient threshold is set for the ground point based on the elevation values ​​of the grid cells in its neighborhood. For each ground point, if the elevation gradient of the ground point is greater than its corresponding elevation gradient threshold, the ground point is determined as a step line feature point that satisfies the elevation gradient constraint condition.

[0072] In embodiments of this application, the processor may also be configured to: The acquired point cloud data of the open-pit mine is subjected to progressive morphological filtering to extract the target point cloud data. For each first ground point in the target point cloud data, the normal vector angle and distance between each second ground point in the neighborhood of the first ground point and the first ground point are determined. For each first ground point, the corresponding first weight and second weight are determined according to the normal vector angle and distance between the first ground point and each second ground point. For each first ground point, bilateral filtering is performed on the second ground points in the neighborhood of the first ground point according to all the first weights and all the second weights corresponding to the first ground point to correct the target point cloud data.

[0073] like Figure 5 As shown in the illustration, this application also provides a bench line detection device for open-pit mines, which may include: The point cloud noise filtering model can be used to acquire point cloud data from open-pit mines and extract target point cloud data, which includes ground points in the point cloud data. It can also be used to perform progressive morphological filtering on the acquired point cloud data from open-pit mines to extract target point cloud data. For each first ground point in the target point cloud data, the normal vector angle and distance between the first ground point and each second ground point in its neighborhood are determined. For each first ground point, corresponding first and second weights are determined based on the normal vector angle and distance between the first and second ground points. For each first ground point, bilateral filtering is performed on the second ground points in its neighborhood based on all the first and second weights corresponding to the first ground point to correct the target point cloud data.

[0074] The dynamic elevation stratification module can be used to perform stratification processing on target point cloud data to generate multiple stratified intervals. It can also be used to obtain the height range of the target point cloud data in the elevation data field; determine the layer thickness of each stratified interval based on the height range and a preset number of stratifications; and generate multiple stratified intervals from the target point cloud data using a sliding window based on the layer thickness and a preset overlap rate. After generating multiple stratified intervals, the slope variation data of each stratified interval is determined; the layer thickness of each stratified interval is adjusted based on the slope variation data; or, for each stratified interval, if the layer thickness of the stratified interval is less than a preset layer thickness threshold, the stratified interval is merged into an adjacent stratified interval, or the stratified interval is allowed to be less than the preset layer thickness threshold.

[0075] The joint step feature discrimination module can be used to select step line feature points from ground points that satisfy the constraints of each dimension based on multiple constraints of different dimensions. Based on the satisfied constraints, the feature point type of each step line feature point is determined, including strong feature points and weak feature points. For each first ground point, the local normal vector of each second ground point in its neighborhood is determined; the angle between each local normal vector and a preset direction (perpendicular to the ground) is determined; for each first ground point, the local normal vectors of all second ground points corresponding to the first ground point are examined to determine if there is a sudden change in the local normal vector of the first ground point within its neighborhood; for any first ground point, if the angle between the local normal vectors corresponding to the first ground point is less than a preset angle threshold, and there is a sudden change in the local normal vector of the first ground point, then the first ground point is determined to be a step line feature point that satisfies the normal vector angle constraint. The moving least squares method is used to fit surfaces to all second ground points within the neighborhood of each first ground point to generate a fitted surface. The Gaussian curvature and mean curvature of each first ground point on the fitted surface are determined. The Gaussian curvature and mean curvature of each first ground point are fused to obtain the composite curvature of each first ground point. For each first ground point, if the composite curvature is less than a preset curvature threshold, the first ground point is determined as a step line feature point satisfying the curvature constraint. The target point cloud data is gridded, and each grid cell is filled with elevation values. For each ground point, an edge detection algorithm is used to determine the elevation gradient of the ground point based on the elevation values ​​of the grid cells within its neighborhood. For each ground point, an elevation gradient threshold is determined based on the elevation values ​​of the grid cells within its neighborhood. For each ground point, if the elevation gradient is greater than its corresponding elevation gradient threshold, the ground point is determined as a step line feature point satisfying the elevation gradient constraint. It can also be used to determine the weight value corresponding to the constraint condition in each dimension for each step line feature point; and to determine the feature point type of each step line feature point by performing a weighted sum based on all weight values ​​and the constraint conditions satisfied by the step line feature point. Furthermore, it can be used to determine the feature point type of step line feature points that satisfy the constraint conditions in all dimensions as strong feature points; and to determine the feature point type of step line feature points that satisfy the constraint conditions in at least one dimension but not in all dimensions as weak feature points.

[0076] The cross-layer feature fusion module can be used to determine the overlapping region between any two adjacent layer intervals; identify target ground points with different feature point types among the ground points included in each overlapping region; and correct the feature point type of the target ground point based on the feature point types of all ground points in the neighborhood of the target ground point. Specifically, it determines the feature point type of each ground point in the neighborhood of the target ground point; determines the proportion of ground points of each feature point type in the neighborhood of the target ground point; and corrects the feature point type of the target ground point to the feature point type corresponding to the maximum proportion.

[0077] The step line fitting and optimization module can be used to use strong feature points as control points and weak feature points as reference points to perform curve fitting on the strong and weak feature points to obtain the step line information of the open-pit mine.

[0078] This application also provides a machine-readable storage medium storing instructions that cause a machine to perform the above-described method for detecting step lines in open-pit mines.

[0079] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor A01, a network interface A02, memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A04. The database stores bench line detection data for open-pit mines. The network interface A02 communicates with external terminals via a network connection. When the processor A01 executes the computer program B02, it implements a bench line detection method for open-pit mines.

[0080] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0081] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0082] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0083] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0084] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0085] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0086] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0087] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0088] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0089] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for detecting bench lines in open-pit mines, characterized in that, The step line detection method includes: Obtain point cloud data from open-pit mines; Extract target point cloud data from the point cloud data, wherein the target point cloud data includes ground points in the point cloud data; Based on constraints of multiple different dimensions, step line feature points that satisfy the constraints of each dimension are selected from the ground points. The feature point type of each step line feature point is determined based on the satisfied constraints, and the feature point type includes strong feature points and weak feature points. Using the strong feature points as control points and the weak feature points as reference points, curve fitting is performed on the strong feature points and the weak feature points to obtain the step line information of the open-pit mine.

2. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The method further includes: The target point cloud data is processed into layers to generate multiple layered intervals; Determine the overlapping area between any two adjacent hierarchical intervals; Identify the target ground points with different feature point types among the ground points included in each overlapping region; The feature point type of the target ground point is corrected based on the feature point types of all ground points in the neighborhood of the target ground point.

3. The method for detecting bench lines in open-pit mines according to claim 2, characterized in that, The step of performing layered processing on the target point cloud data to generate multiple layered intervals includes: Obtain the height range of the target point cloud data in the elevation data field; The layer thickness of each layer interval is determined based on the height range and the preset number of layers; Based on the layer thickness and preset overlap rate, a sliding window is used to generate multiple layered intervals from the target point cloud data.

4. The method for detecting bench lines in open-pit mines according to claim 3, characterized in that, The method further includes: After generating multiple stratified intervals, determine the slope variation data for each stratified interval; Adjust the layer thickness of each layer interval according to the slope change data; or For each layered interval, if the layer thickness of the layered interval is less than a preset layer thickness threshold, the layered interval is merged into a layered interval adjacent to the layered interval, or the layered interval is allowed to be less than the preset layer thickness threshold.

5. The method for detecting bench lines in open-pit mines according to claim 2, characterized in that, The step of correcting the feature point type of the target ground point based on the feature point types of all ground points in the neighborhood of the target ground point includes: Determine the feature point type of each ground point within the neighborhood of the target ground point; Determine the proportion of ground points of each feature point type in the neighborhood of the target ground point; The feature point type of the target ground point is corrected to the feature point type corresponding to the maximum value of the percentage.

6. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The feature point type determined based on the satisfied constraints includes: The feature point type of the step line feature points that satisfy the constraints of all dimensions is determined as the strong feature points. The feature point type of a step line feature point that satisfies at least one dimension of the constraint conditions but not all dimensions of the constraint conditions is determined as a weak feature point.

7. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The feature point type determined based on the satisfied constraints includes: For each step line feature point, determine the weight value corresponding to the constraint condition in each dimension; For each step line feature point, a weighted sum is performed based on all weight values ​​and the constraints satisfied by the step line feature point to determine the feature point type of the step line feature point.

8. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The step line feature points selected from the ground points based on constraints of multiple different dimensions, each satisfying the constraints of a specific dimension, include: For each first ground point, determine the local normal vector of each second ground point within the neighborhood of the first ground point; Determine the angle between each local normal vector and a preset direction, where the preset direction is perpendicular to the ground. For each first ground point, determine whether there is a sudden change in the local normal vector of the first ground point within the neighborhood of the first ground point, based on the local normal vectors of all second ground points corresponding to the first ground point. For any first ground point, if the angle between the normal vectors corresponding to the first ground point is less than a preset angle threshold, and there is a sudden change in the local normal vector of the first ground point, the first ground point is determined to be a step line feature point that satisfies the normal vector angle constraint condition.

9. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The step line feature points selected from the ground points based on constraints of multiple different dimensions, each satisfying the constraints of a specific dimension, include: The moving least squares method is used to fit a surface to all second ground points in the neighborhood of each first ground point to generate a fitted surface. Determine the Gaussian curvature and mean curvature of the fitted surface corresponding to each first ground point; The Gaussian curvature and the mean curvature corresponding to each first ground point are fused to obtain the composite curvature of each first ground point; For each first ground point, if the composite curvature of the first ground point is less than a preset curvature threshold, the first ground point is determined to be a step line feature point that satisfies the curvature constraint condition.

10. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The step line feature points selected from the ground points based on constraints of multiple different dimensions, each satisfying the constraints of a specific dimension, include: The target point cloud data is processed into a grid, and each grid cell is filled with an elevation value; For each ground point, the elevation gradient of the ground point is determined based on the elevation values ​​of the grid cells included in the neighborhood of the ground point using an edge detection algorithm. For each ground point, an elevation gradient threshold is determined based on the elevation values ​​of the grid cells within the neighborhood of the ground point. For each ground point, if the elevation gradient of the ground point is greater than its corresponding elevation gradient threshold, the ground point is determined to be a step line feature point that satisfies the elevation gradient constraint condition.

11. The method for detecting bench lines in open-pit mines according to claim 1, characterized in that, The target point cloud data extracted from the point cloud data includes: The point cloud data of the acquired open-pit mine is subjected to progressive morphological filtering to extract the target point cloud data. For each first ground point in the target point cloud data, determine the angle between the normal vectors of each second ground point in the neighborhood of the first ground point and the first ground point, as well as the distance between them; For each first ground point, the corresponding first weight and second weight are determined based on the angle between the normal vectors of the first ground point and each second ground point and the distance between them. For each first ground point, bilateral filtering is performed on the second ground points in the neighborhood of the first ground point according to all first weights and all second weights corresponding to the first ground point, so as to correct the target point cloud data.

12. A bench line detection device for open-pit mines, characterized in that, include: The memory is configured to store instructions; The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the step line detection method for open-pit mines according to any one of claims 1 to 11.

13. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the bench line detection method for open-pit mines according to any one of claims 1 to 11.

Images