A method for extracting an under-segmented point cloud of a blast pile

Abstract

The present application belongs to the technical field of underground metal mine blasting engineering, and relates to a kind of blast pile point cloud under segmentation extraction method.The method comprises: obtaining blast pile original three-dimensional point cloud data;The blast pile original three-dimensional point cloud data is preprocessed, and the blast pile three-dimensional point cloud model is reconstructed;Based on KD tree, the neighborhood search structure is constructed and the resolution is determined;The blast pile three-dimensional point cloud model is segmented to generate segmented result point cloud;Through principal component analysis, an oriented bounding box (OBB) is constructed, its orthogonal direction vector and edge length are obtained;The OBB direction vector and the plane are used as the projection direction and the plane, and three orthogonal projection point clouds are obtained;According to the projection point cloud parameter, the neighborhood radius is set, and the local density is calculated;The maximum value of the density is counted, and the point cloud greater than the threshold value is determined as the under-segmentation point cloud, which needs to be re-segmented.The beneficial effects are: through the local density analysis of the plane projection point cloud, the under-segmentation point cloud is accurately identified, which provides a basis for adjusting the segmentation strategy and improves the segmentation accuracy and precision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of underground metal mine blasting engineering, and particularly relates to a method for extracting under-segmented point clouds of a blasting pile. BACKGROUND

[0002] In underground metal mine blasting, three-dimensional scanning technology has become an important tool for reconstructing the three-dimensional morphology of a blasting pile. Based on the point cloud data of the blasting pile obtained by three-dimensional scanning, an efficient and accurate point cloud segmentation method is used to segment the point cloud of the blasting pile, which can provide accurate information for blasting pile size measurement. Due to the significant differences in morphology and size of the blasting pile rock, the obtained point cloud data of the blasting pile usually presents irregular morphology, large density difference and complex stacking characteristics.

[0003] Common point cloud segmentation methods, such as region growing algorithm, DBSCAN algorithm and VCCS+LCCP algorithm, usually rely on geometric features or set rules for segmentation. The high complexity of the point cloud of the blasting pile makes these segmentation methods have low precision and poor completeness, especially when dealing with point clouds of the blasting pile with large regional density differences, which is prone to under-segmentation.

[0004] In addition, the multi-scale characteristics of the blasting pile rock make it difficult for a single scale segmentation method to comprehensively and accurately identify the blasting pile rock. SUMMARY

[0005] Technical problems to be solved

[0006] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides a method for extracting under-segmented point clouds of a blasting pile, which aims to solve the problem of low precision and poor completeness of existing segmentation methods caused by the high complexity of the point cloud of the blasting pile, especially when dealing with point clouds of the blasting pile with large regional density differences, extracting under-segmented point clouds under multiple scales for subsequent re-segmentation, which can effectively avoid under-segmentation.

[0007] Technical scheme

[0008] In order to achieve the above-mentioned purpose, the main technical scheme adopted by the present application comprises:

[0009] In a first aspect, the present application provides a method for extracting under-segmented point clouds of a blasting pile, comprising the following steps:

[0010] Step 1: obtaining original three-dimensional point cloud data of the blasting pile;

[0011] Step 2: preprocessing the original three-dimensional point cloud data of the blasting pile to reconstruct a three-dimensional point cloud model of the blasting pile;

[0012] Step 3: constructing a point cloud space neighborhood search structure based on a KD tree and determining a point cloud resolution;

[0013] Step 4: Segment the 3D point cloud model of the burst pile and generate the segmentation result point cloud;

[0014] Step 5: For each segmented point cloud, principal component analysis is used to determine the principal direction of the segmented point cloud, and an oriented bounding box (OBB) is constructed based on the principal direction to obtain the three mutually orthogonal direction vectors of the oriented bounding box and their side lengths in the corresponding directions.

[0015] Step 6: For each segmented point cloud, take the three orthogonal direction vectors of its bounding box and the corresponding bounding box plane as the projection direction and projection plane, respectively, and perform orthogonal projection to obtain three orthogonal projected point clouds. ;

[0016] Step 7: Based on the projected point cloud The neighborhood search radius is set by the side length of the bounding box plane and the point cloud resolution, and the local density of the point cloud is calculated based on the neighborhood search radius;

[0017] Step 8: Statistical segmentation result point cloud Planar projection point cloud Local density maximum R max Compare it with the set density threshold R th The comparison is used to determine whether the segmentation is complete.

[0018] As a further improvement to the method of the present invention, step 2 specifically includes the following method:

[0019] Step 2.1: Use the SOR algorithm to denoise the original 3D point cloud data of the burst pile, remove outliers and edge noise in the original 3D point cloud data of the burst pile, and obtain the denoised 3D point cloud data of the burst pile.

[0020] Step 2.2: Based on the voxel downsampling algorithm, downsample the denoised 3D point cloud data of the burst pile to reconstruct the 3D point cloud model of the burst pile.

[0021] As a further improvement to the method of the present invention, step 3 involves constructing a KD tree and determining the point cloud resolution using the average neighborhood distance, specifically including the following methods:

[0022] Step 3.1: Based on the three-dimensional spatial coordinates of each point in the point cloud, the KD tree algorithm is used to spatially divide the point cloud and establish the spatial neighborhood search structure of the point cloud, i.e., the KD tree structure.

[0023] Step 3.2: Using any point in the point cloud p i For the query point, perform a neighborhood search in the KD-tree structure to obtain the point. p ia neighborhood point set within a preset spatial range;

[0024] Step 3.3: based on the spatial distance between points in the neighborhood point set and the query point, calculate the average neighborhood distance of the point cloud, and take the average neighborhood distance as the point cloud resolution .

[0025] As a further improvement of the method of the application, in step 5, specifically: for each segmented result point cloud , first calculate its covariance matrix, and perform eigenvalue decomposition on the covariance matrix to obtain a set of eigenvectors and corresponding eigenvalues : ; wherein the eigenvector corresponds to the principal direction of the point cloud, and the eigenvectors constitute a principal direction set of the point cloud in three-dimensional space, and the principal direction set is composed of three mutually orthogonal direction vectors ;

[0026] Based on the principal direction , construct a direction matrix of the oriented bounding box, and take the centroid of the segmented result point cloud as the origin of the local coordinate system; map the coordinates of all points in the segmented result point cloud to the local coordinate system defined by the direction matrix and the origin;

[0027] For any one principal direction , the projection of any point in the segmented result point cloud on the principal direction is , and the extreme range of on each principal direction is determined in the coordinate system , thereby obtaining the edge length of the oriented bounding box in three orthogonal directions .

[0028] As a further improvement of the method of the application, in step 7, specifically comprising the following method:

[0029] Step 7.1: set the neighborhood search radius according to the edge length of the bounding box plane where the projected point cloud is located and the point cloud resolution ;

[0030] Step 7.2: calculate the local density of the projected point cloud according to the neighborhood search radius of the projected point cloud .

[0031] As a further improvement of the method of the application, in step 7.1, specifically: according to the edge length of the bounding box plane where the projected point cloud is located and the point cloud resolution ( ), calculate the projected point cloud The length of the diagonal of the bounding box plane At the same time, based on the determined point cloud resolution Determine the projected point cloud neighborhood search radius .

[0032] As a further improvement to the method of the present invention, step 7.2 specifically involves: projecting the point cloud. any point in space ,by Centered on the neighborhood search radius Determine the midpoint of the point cloud. neighborhood point set Calculate the neighboring points in the neighborhood point set. Between the current spatial point European distance The Gaussian kernel function is used to perform distance-weighted accumulation on neighboring points to calculate the projected point cloud. Local density .

[0033] As a further improvement to the method of the present invention, in step 8, if R max >R th If R is not segmented, it is determined to be an under-segmented point cloud, which needs to be extracted and re-segmented; max ≤R th If the segmentation is complete, there are no unsegmented point clouds.

[0034] In a second aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed, implements the method for undersegmentation extraction of burst point clouds as described in any of the first aspects above.

[0035] Thirdly, the present invention provides a storage device, including a storage medium and a processor, wherein the storage medium stores a computer program, and when the program is executed by the processor, it implements the method for undersegmentation extraction of burst point clouds as described in any of the first aspects above.

[0036] The beneficial effects of this invention are:

[0037] 1. Realize the objective and automatic judgment of segmentation integrity. The invention overcomes the subjective limitations of traditional methods relying on artificial experience, and innovatively proposes an analysis framework based on multi-orthogonal projection density fusion. First, the main direction of the segmentation result point cloud is determined by principal component analysis and a directed bounding box is constructed, then two-dimensional point clouds are obtained by projecting along three orthogonal directions, and finally a comprehensive judgment is made based on the local density distribution characteristics of the projected point cloud. This method avoids the one-sidedness of a single criterion through multi-view complementation, combines an adaptive parameter adjustment mechanism with a quantitative point cloud density standard, and realizes the objective and automatic judgment of segmentation integrity.

[0038] 2. The feature has clear physical meaning and efficient calculation. The three-dimensional under-segmentation problem is transformed into two-dimensional density analysis through dimensionality reduction by projection, realizing the unity of clear physical meaning and efficient calculation. The local density of the projected point cloud directly reflects the geometric shape of the rock mass: the projection density distribution of a single complete rock mass is uniform, and the under-segmented rock mass appears abnormal aggregation, which guarantees the strong interpretability of the results. At the same time, the three-dimensional neighborhood search is converted to two-dimensional calculation, effectively reducing the time complexity and improving the processing efficiency, meeting the real-time demand of engineering.

[0039] 3. Wide compatibility and strong adaptability with existing technologies. As a general post-segmentation processing module, it can be seamlessly integrated with various point cloud segmentation algorithms, and through the adaptive parameter mechanism, it can adapt to different acquisition conditions and rock mass sizes, having good engineering applicability and promotional value. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 A process flow chart of an under-segmentation extraction method for blast pile point cloud is provided for the embodiments of the invention.

[0041] Figure 2 The original three-dimensional point cloud data of the blast pile in the embodiments of the invention;

[0042] Figure 3 The three-dimensional point cloud model of the blast pile in the embodiments of the invention;

[0043] Figure 4 The result graph after the three-dimensional point cloud model of the blast pile is segmented by the DBSCAN algorithm in the embodiments of the invention;

[0044] Figure 5 The projected point cloud graph of a certain segmentation result of the three-dimensional point cloud model of the blast pile in the embodiments of the invention on its OBB three orthogonal planes;

[0045] Figure 6 The projected point cloud density distribution statistical graph of the segmentation result of the three-dimensional point cloud model of the blast pile in the embodiments of the invention;

[0046] Figure 7The under-segmentation result point cloud of the blast pile three-dimensional point cloud model extracted according to the point cloud density threshold in the embodiment of the present application;

[0047] Figure 8 The under-segmentation result point cloud of the blast pile three-dimensional point cloud model extracted according to the point cloud density threshold in the embodiment of the present application; DETAILED DESCRIPTION

[0048] In order to better explain the present application, so as to be understood, the present application is described in detail by specific embodiments in combination with the accompanying drawings. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided so that the present application can be more clearly, thoroughly understood, and the scope of the present application can be completely conveyed to those skilled in the art.

[0049] In a first aspect, as Figure 1 indicated, the present application provides a blast pile point cloud under-segmentation extraction method, comprising the following steps:

[0050] Step 1: obtaining the original three-dimensional point cloud data of the blast pile by using a three-dimensional scanner;

[0051] Specifically, in step 1, a handheld three-dimensional scanner is used, scanning parameters are set according to the characteristics of the scanner and the shape of the rock pile, a scanning path is planned, and multi-angle and multi-position scanning is performed; during the scanning process, each scanning data is automatically spliced, and finally the original three-dimensional point cloud data of the blast pile is obtained, as shown in Figure 2 .

[0052] Step 2: pre-processing the original three-dimensional point cloud data of the blast pile and reconstructing the three-dimensional point cloud model of the blast pile; the pre-processing includes point cloud denoising and down-sampling;

[0053] Step 2.1: using SOR algorithm to denoise the original three-dimensional point cloud data of the blast pile, removing outliers and edge noise points in the original three-dimensional point cloud data of the blast pile, and obtaining denoised three-dimensional point cloud data of the blast pile;

[0054] Step 2.2: based on the voxel down-sampling algorithm, the denoised three-dimensional point cloud data of the blast pile is down-sampled to reconstruct the three-dimensional point cloud model of the blast pile;

[0055] The three-dimensional point cloud model of the blast pile in this embodiment is shown in Figure 3 .

[0056] Step 3: constructing a point cloud space neighborhood search structure based on KD tree (k-d tree), supporting fast neighborhood search operation of any point in the point cloud, and determining the point cloud resolution;

[0057] Step 3.1: Based on the three-dimensional spatial coordinates of each point in the point cloud, the KD tree algorithm is used to spatially divide the point cloud and establish the spatial neighborhood search structure of the point cloud, i.e., the KD tree structure.

[0058] Step 3.2: Using any point in the point cloud p i For the query point, perform a neighborhood search in the KD-tree structure to obtain the point. p i The set of neighboring points within a preset spatial range;

[0059] Step 3.3: Based on the spatial distance between the neighborhood point set point and the query point, calculate the average neighborhood distance of the point cloud, and use this average neighborhood distance as the point cloud resolution. .

[0060] Step 4: Use the DBSCAN segmentation algorithm to segment the reconstructed 3D point cloud model of the burst pile and generate the segmentation result point cloud.

[0061] In this embodiment, a total of 77 segmentation result point clouds were generated, and the segmentation results are as follows: Figure 4 As shown.

[0062] Step 5: For each segmented point cloud, Principal Component Analysis (PCA) is used to determine the principal direction of the segmented point cloud, and a directed bounding box is constructed based on the principal direction to obtain the three mutually orthogonal direction vectors of the directed bounding box and their side lengths in the corresponding directions.

[0063] The directed bounding box consists of three pairwise orthogonal direction vectors, which correspond to the first principal direction, the second principal direction, and the third principal direction of the point cloud in space, respectively. The three direction vectors together define the local coordinate system of the point cloud.

[0064] Specifically, for each segmentation result point cloud First, its covariance matrix is ​​calculated, and then eigenvalue decomposition is performed on the covariance matrix to obtain a set of eigenvectors. and the corresponding eigenvalues : Among them, the feature vector Corresponding to the principal directions of the point cloud, the feature vectors constitute the principal direction group of the point cloud in three-dimensional space. The principal direction group consists of three pairwise orthogonal direction vectors. composition.

[0065] Based on the main direction Construct a direction matrix for a directed bounding box, using the centroid of the segmented point cloud as the origin of the local coordinate system; map the coordinates of all points in the segmented point cloud to the local coordinate system defined by the direction matrix and the origin;

[0066] for any one of the principal directions , the segmented point cloud any point in the projection in the principal direction , determine the extreme range of the point cloud in each principal direction under the coordinate system , so as to obtain the edge length of the oriented bounding box in three orthogonal directions .

[0067] Step 6: For each segmented point cloud, take the three orthogonal direction vectors of its bounding box and the corresponding bounding box planes as the projection direction and projection plane respectively, perform orthogonal projection to obtain three orthogonal projection point clouds .

[0068] In this embodiment, steps 5 and 6 are performed on any segmentation result generated in step 4 to obtain three orthogonal plane projection point clouds of the segmentation result point cloud, i.e., plane projection point cloud 1, plane projection point cloud 2 and plane projection point cloud 3. The obtained plane projection point clouds are as shown in Figure 5 .

[0069] Step 7: Set the neighborhood search radius according to the edge length of the bounding box plane where the projection point cloud is located and the point cloud resolution, and calculate the local density of the point cloud based on the neighborhood search radius.

[0070] Step 7.1: Set the neighborhood search radius according to the edge length of the bounding box plane where the projection point cloud is located and the point cloud resolution .

[0071] Specifically, according to the edge length of the bounding box plane where the projection point cloud is located and ( ), the diagonal length of the bounding box plane where the projection point cloud is located is calculated , and according to the determined point cloud resolution , the neighborhood search radius of the projection point cloud is determined.

[0072] Step 7.2: Calculate the local density of the projection point cloud according to the neighborhood search radius of the projection point cloud .

[0073] Specifically, for any spatial point in the projection point cloud , take as the center, and according to the neighborhood search radius , calculate the number of points Determine the midpoint of the point cloud. neighborhood point set Calculate the neighboring points in the neighborhood point set. Between the current spatial point European distance The Gaussian kernel function is used to perform distance-weighted accumulation on neighboring points to calculate the projected point cloud. Local density .

[0074] Step 8: Calculate the local density maximum R of the planar projection point cloud of the segmented point cloud. max Compare it with the set density threshold R th Compare the results to determine whether the segmentation is complete;

[0075] If R max >R th If the value is not specified, it is determined to be an under-segmented point cloud, which needs to be extracted and re-segmented.

[0076] If R max ≤R th If the segmentation is complete, there are no unsegmented point clouds.

[0077] Specifically, according to steps 6 and 7, the local density features of the projected point cloud of the segmented point cloud on its three oriented bounding box planes are calculated, and the maximum local density is used as the density characterization parameter of the segmented point cloud.

[0078] like Figure 6 As shown, the above calculation process is repeated for all segmented point clouds to obtain the distribution of local density maximum values ​​for each segmented point cloud. Based on this distribution, a local density threshold R is set. th =220. When the maximum local density of the segmented point cloud is lower than this threshold, the segmented point cloud is determined to be an under-segmented point cloud.

[0079] The segmentation point cloud extracted based on the above threshold determination rules is as follows: Figure 7 As shown.

[0080] In this embodiment, taking one segmented point cloud as an example, the maximum local density of its three projected point clouds is 1649.10, and this is compared with the actual under-segmented point cloud (such as...). Figure 8 As shown in the figure, the undersegmentation extraction method for the burst pile point cloud achieves an accuracy rate of 93% in identifying undersegmented point clouds. The accuracy rate of the undersegmented point cloud identification is calculated as the percentage of the number of undersegmented result point clouds identified by the present invention to the actual number of undersegmented result point clouds.

[0081] In the embodiment, the DBSCAN segmentation algorithm used, the number of generated segmentation result point clouds, the calculation method of local density, and the specific parameters such as the density threshold are only exemplary settings for illustrating the technical solutions of the present application and do not constitute a limitation on the present application.

[0082] In a second aspect, the embodiments of the present application provide a computer readable storage medium having a computer program stored thereon, the program being executed to implement the point cloud under-segmentation extraction method of the first aspect.

[0083] In a third aspect, the embodiments of the present application provide a storage device including a storage medium and a processor, the storage medium storing a computer program, the program being executed by the processor to implement the point cloud under-segmentation extraction method of any one of the first aspect.

[0084] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0085] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application should also include these modifications and variations.

[0086] Although the embodiments of the present application have been shown and described above, it should be understood that the above-described embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-described embodiments within the scope of the present application.

Claims

1. A method for extracting an under-segmented blast pile point cloud, characterized in that, The method comprises the following steps: Step 1: obtaining original three-dimensional point cloud data of a blast pile; Step 2: preprocessing the original three-dimensional point cloud data of the blast pile and reconstructing a three-dimensional point cloud model of the blast pile; Step 3: constructing a point cloud space neighborhood search structure based on a KD tree and determining a point cloud resolution; Step 4: segmenting the three-dimensional point cloud model of the blast pile to generate segmented result point clouds; Step 5: for each segmented result point cloud, determining a main direction of the segmented result point cloud by principal component analysis and constructing an oriented bounding box (OBB) based on the main direction to obtain three mutually orthogonal direction vectors of the OBB and edge lengths of the OBB in the corresponding directions; Step 6: For each segmented result point cloud, perform orthogonal projection with three orthogonal direction vectors of its bounding box and corresponding bounding box planes as projection direction and projection plane respectively, to obtain three orthogonal projection point clouds ; Step 7: According to the projected point cloud The length of the bounding box plane and the point cloud resolution set the neighborhood search radius, and the point cloud local density is calculated based on the neighborhood search radius; The projection point cloud The length of the bounding box plane and the point cloud resolution set the neighborhood search radius, specifically: According to the projection point cloud Side length of the bounding box plane and ( ), calculate the projected point cloud The length of the diagonal of the bounding box plane At the same time, based on the determined point cloud resolution Determine the projected point cloud neighborhood search radius ; The local density of the point cloud is calculated based on a neighborhood search radius, and the calculation method is as follows: For projected point clouds any point in space ,by Centered on the neighborhood search radius Determine the midpoint of the point cloud. neighborhood point set Calculate the neighboring points in the neighborhood point set. Between the current spatial point European distance The Gaussian kernel function is used to perform distance-weighted accumulation on neighboring points to calculate the projected point cloud. Local density ; Step 8: Counting the segmented point cloud the planar projection point cloud the local density maximum R max and compare it with the set density threshold R th to determine whether the segmentation is complete.

2. The method of claim 1, wherein, In step 2, the method specifically comprises the following steps: Step 2.1: denoising the original three-dimensional point cloud data of the blast pile by using a SOR algorithm, removing outliers and edge noise points in the original three-dimensional point cloud data of the blast pile, and obtaining denoised three-dimensional point cloud data of the blast pile; Step 2.2: performing down-sampling on the denoised three-dimensional point cloud data of the blast pile based on a voxel down-sampling algorithm to reconstruct the three-dimensional point cloud model of the blast pile.

3. The method of claim 2, wherein, In step 3, the KD tree is constructed, and the point cloud resolution is determined by the average neighborhood distance, and the method specifically comprises the following steps: Step 3.1: according to the three-dimensional space coordinates of each point in the point cloud, the point cloud is divided in space by using a KD tree algorithm to establish a point cloud space neighborhood search structure, i.e. a KD tree structure; Step 3.2: For any point p in the point cloud i perform a neighborhood search in the KD-tree structure for the query point, obtaining the point p i neighborhood point set within a preset spatial range; Step 3.3: Calculate the average neighborhood distance of the point cloud based on the spatial distance between points in the neighborhood set and the query point, and take the average neighborhood distance as the resolution of the point cloud .

4. The method of claim 3, wherein, In step 5, specifically: for each segmented result point cloud , first calculate its covariance matrix, and perform eigenvalue decomposition on the covariance matrix to obtain a set of eigenvectors and corresponding eigenvalues : ; wherein the eigenvector corresponds to the principal direction of the point cloud, and the eigenvectors constitute the principal direction set of the point cloud in the three-dimensional space, and the principal direction set is composed of three mutually orthogonal direction vectors ; based on the main direction A direction matrix of the oriented bounding box is constructed, and the segmentation result point cloud is mapped into a local coordinate system defined by the direction matrix. For any one main direction , the projection of the segmentation result point cloud in any point in this direction is , the extreme range of the point cloud in each main direction is determined under this coordinate system , so as to obtain the edge length of the oriented bounding box in three orthogonal directions .

5. The method of claim 4, wherein, In step 8, if R max > R th , it is determined that there is an under-segmented point cloud, and the point cloud needs to be extracted and re-segmented; if R max ≤ R th , it is determined that the segmentation is completed, and there is no under-segmented point cloud.

6. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the blast pile point cloud under-segmentation extraction method in any one of claims 1 to 5.

7. A storage device comprising a storage medium and a processor, the storage medium storing a computer program, characterized in that, The processor executes the computer program to realize the blast pile point cloud under-segmentation extraction method in any one of claims 1 to 5.

Images