A rock mass dangerous easy-to-slide three-dimensional point cloud block automatic identification method

Abstract

The application provides a rock mass dangerous easy-to-slide three-dimensional point cloud block automatic identification method, and belongs to the technical field of mining engineering. The method first uses a three-dimensional Gaussian mixture model algorithm to automatically identify rock mass surface structural planes, and obtains each classification cluster in each group of rock mass surface structural planes. Each classification cluster in all rock mass surface structural planes is stored in a specified index mode. Then, adjacent clusters of each classification cluster in all rock mass surface structural planes are found, the vertices of rock mass candidate blocks are determined, and a block growing algorithm is used to automatically identify rock mass candidate blocks. Finally, the spatial topological relationship between the rock mass candidate blocks and the surrounding rock mass candidate blocks or rock mass surface structures is judged, so as to judge the dangerous blocks, that is, the dangerous easy-to-slide three-dimensional point cloud blocks. The method can be applied to the identification of rock mass slope dangerous easy-to-slide three-dimensional point cloud blocks in complex scenes such as mines in all directions. The method is simple to realize, has remarkable effects, and meets the application requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of mining engineering, and in particular to a method for automatically identifying three-dimensional point cloud blocks of dangerous and easily sliding rock masses. Background Art

[0002] Rock slopes contain natural fractures, folds, cleavages, and other linear, planar, and volumetric geological structures. These structures, under the influence of external stresses, can easily lead to structural instability, resulting in landslides or collapses. This can severely damage the natural ecological environment and even threaten human life and property. Therefore, the presence of dangerous, easily sliding three-dimensional blocks in rock slopes is a significant concern for construction workers and engineers.

[0003] Traditionally, volume identification and information acquisition typically involves contact measurement, where surveyors use handheld tools such as geological compasses, measuring ropes, and rulers to obtain metrics such as the rock mass's surface attitude, spacing, and trace length. However, in practice, these traditional measurement methods can produce significant errors due to the complexity of both the surveyor and the site conditions. Furthermore, they are time-consuming, labor-intensive, and dangerous. With the increasing scale and speed of modern rock mass engineering projects, traditional measurement methods are unable to meet these demands. Furthermore, for complex construction sites such as steep slopes, mine slopes, and island slopes, traditional measurement methods make it difficult for surveyors to reach the work area. With the rapid advancement of measurement technology, non-contact measurement methods, such as drones, 3D laser scanners, and digital cameras, have become increasingly popular among researchers. These methods can rapidly acquire rock mass characterization data without contacting the rock mass surface. Compared to traditional methods, these methods are more time-efficient, labor-intensive, and less hazardous, making them suitable for safely and efficiently acquiring raw information on structural surfaces in modern large-scale geotechnical engineering projects.

[0004] To date, researchers have conducted limited research on block identification, using methods such as flood fill, block search, alpha shape algorithms, and depth-first algorithms. Some researchers have also manually delineated rock mass blocks using various processing software. Existing research typically only identifies regular hexahedrons and irregular blocks, but not all of these blocks are dangerous and prone to landslides. Most of these blocks pose no risk of landslide and have no direct impact on rock slope landslides. However, the number of dangerous and prone blocks that could potentially impact landslides is small and difficult to identify. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to address the deficiencies of the above-mentioned existing technologies and provide a method for automatically identifying dangerous and prone to sliding three-dimensional point cloud blocks of rock masses. The method uses a block growth algorithm to automatically identify candidate rock masses and judge and identify dangerous and prone to sliding three-dimensional point cloud blocks.

[0006] To solve the above technical problems, the present invention adopts a technical solution: a method for automatically identifying dangerous and prone to sliding three-dimensional point cloud blocks of rock mass, comprising the following steps:

[0007] Step 1: Use the three-dimensional Gaussian mixture model algorithm to automatically identify the rock surface structural planes, and obtain each classification cluster in each group of rock surface structural planes, and store each classification cluster in all rock surface structural planes in a specified indexing manner;

[0008] Step 2: Find the adjacent clusters of each classification cluster in all rock surface structural surfaces, determine the vertices of the rock candidate blocks, and automatically identify the rock candidate blocks using the block growing algorithm, including the following steps:

[0009] Step 2.1: Find the adjacent clusters of each classification cluster in all rock surface structural surfaces and store them in the adjacent cluster list of the search cluster;

[0010] Take one of the clusters To start the search cluster, use the neighborhood search method to traverse all the remaining classification clusters and obtain the adjacent clusters of all the remaining classification clusters List of adjacent clusters that coexist with the search cluster Where i is the serial number of the rock surface structural surface, j is the serial number of the cluster in each group of rock surface structural surfaces;

[0011] Step 2.2: Find three adjacent classification clusters, determine the vertices of the rock mass candidate blocks, and store them in the rock mass candidate block vertex list verticesList;

[0012] Find three adjacent classification clusters, and select one of them To start the search cluster, traverse the cluster The list of adjacent clusters of the search cluster Simultaneously search the adjacent cluster list The list of adjacent clusters for each cluster in The intersection points of three adjacent classification clusters are used as vertices of the candidate rock mass blocks. In the same way, the vertices of all candidate rock mass blocks are obtained and stored in the candidate rock mass block vertex list verticesList.

[0013] Step 2.3: Using the block growing algorithm to identify candidate rock blocks, including the following steps:

[0014] Step 2.3.1: Establish the initial space of the candidate rock mass block, calculate the opening direction of the initial space of the candidate rock mass block, and compare the angle between the opening direction and the normal direction with 90 degrees to determine whether the initial state of the candidate rock mass block meets the conditions; index the other vertices on the edge line where the vertex that meets the initial state of the candidate rock mass block is located, and execute step 2.3.2 for the next judgment;

[0015] Calculate the inclination, dip and normal direction of the original rock mass surface, and specify the normal direction V of the original rock mass surface n Towards the inside of the rock mass surface; the rock mass candidate block initially grows from a vertex, and the first vertex is formed by the intersection of three rock mass surfaces. The initial three surfaces can be used to determine a unique space, namely the initial space of the rock mass candidate block;

[0016] Calculate the opening direction V of the initial space of the rock mass candidate block i When V i With V n When the angle is greater than 90 degrees, it means that this vertex is not a valid starting point and does not meet the starting state of the candidate rock mass block, and the next intersection point is determined; when V i With V n When the angle is less than 90 degrees, it means that this vertex is a valid starting point and meets the starting state of the candidate rock mass block;

[0017] Add the vertices that meet the initial state of the rock candidate block to the vertex list verticesList of the rock candidate block, add the first cluster to the planesList of the rock candidate block surface, and index the other vertices on the edge line where the vertex that meets the initial state of the rock candidate block is located, and execute step 2.3.2 for the next judgment; the edge line is the intersection line of the two rock surface surfaces;

[0018] Step 2.3.2: Determine whether the vertex on the endpoint of the edge line is the endpoint of the edge line. If the vertex is on the endpoint of the edge line, execute step 2.3.3 for the next step of determination;

[0019] Each edge line contains multiple vertices. If the vertex is inside the edge line, determine the next vertex in the vertex list verticesList of the rock mass candidate block; if the vertex is at the endpoint of the edge line, execute step 2.3.3 to make the next judgment;

[0020] Step 2.3.3: Determine whether the vertex at the edge line endpoint is on the opening side of the rock mass candidate block. If so, add the vertex at the edge line endpoint to the rock mass candidate block vertex list verticesList;

[0021] If the vertex at the edge line endpoint is outside the rock mass candidate block, then the vertex at the edge line endpoint does not belong to the rock mass candidate block, and the next vertex is determined; if the vertex at the edge line endpoint is inside the rock mass candidate block, then the vertex at the edge line endpoint belongs to the candidate block, and the vertex at the edge line endpoint is added to the rock mass candidate block vertex list verticesList, and the other vertices on the edge line where the vertex is located are indexed, and step 2.3.4 is executed for the next step of determination;

[0022] Step 2.3.4: Determine whether the cluster contained in the newly added vertex already exists in the candidate block. If not, add the new cluster to the rock candidate block surface list planesList;

[0023] Each vertex contains three clusters. Check whether the three clusters corresponding to the vertex at the edge endpoint already exist in the rock candidate block list. If not, add the new cluster to the rock candidate block surface list planesList, and the number of clusters k becomes k+1. Repeat this process, traverse all clusters until all candidate blocks are indexed, then execute step 2.3.5 to proceed to the next step of judgment. The rock candidate block list includes the rock candidate block surface list planesList and the rock candidate block vertex list verticesList.

[0024] Step 2.3.5: Determine whether there are duplicate blocks in the candidate rock mass blocks. If so, delete the duplicate blocks. Otherwise, go to step 3.

[0025] At this time, the rock candidate block consists of the vertex list verticesList and the surface list planesList of the rock candidate block. The number of vertices and the order of vertices in the vertex list verticesList of the rock candidate block are judged. If the number of vertices is the same but the order is different, the rock candidate block is retained, and the number of rock candidate blocks m is increased to m+1, until all rock candidate blocks are judged.

[0026] Step 3: Determine the spatial topological relationship between the candidate rock mass block and the surrounding candidate rock mass blocks or the rock mass surface structure, thereby determining the dangerous block, that is, the dangerous and prone to sliding 3D point cloud block;

[0027] Index the lower surface of each candidate block; if the lower surface of a candidate block does not intersect with any position on the rock surface, mark this rock candidate block as a dangerous block; and so on, judge all rock candidate blocks, mark all dangerous blocks, and all marked dangerous blocks are called dangerous and prone to sliding 3D point cloud blocks.

[0028] The beneficial effects of adopting the above technical solution are: the present invention provides a method for automatically identifying three-dimensional point cloud blocks of dangerous and prone to sliding rock masses, which can automatically identify three-dimensional point cloud blocks of dangerous and prone to sliding rock masses, and can be fully applicable to the identification of three-dimensional point cloud blocks of dangerous and prone to sliding rock slopes in complex scenes such as mines; the method of the present invention is simple to implement, has significant effects, and meets the application requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A flow chart of a method for automatically identifying dangerous and prone-to-sliding three-dimensional point cloud blocks of rock masses provided by an embodiment of the present invention;

[0030] Figure 2 Flowchart of a block growth algorithm provided by an embodiment of the present invention;

[0031] Figure 3 A process diagram of identifying candidate rock blocks using a block growth algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0033] In this embodiment, a method for automatically identifying dangerous and prone to sliding three-dimensional point cloud blocks of rock mass is provided. Figure 1 As shown, the following steps are included:

[0034] Step 1: Use the three-dimensional Gaussian mixture model algorithm to automatically identify the rock surface structural planes, and obtain each classification cluster in each group of rock surface structural planes, and store each classification cluster in all rock surface structural planes in a specified indexing manner;

[0035] Step 2: Find the adjacent clusters of each classification cluster in all rock surface structural surfaces, determine the vertices of the rock candidate blocks, and automatically identify the rock candidate blocks using the block growing algorithm, including the following steps:

[0036] Step 2.1: Find the adjacent clusters of each classification cluster in all rock surface structural surfaces and store them in the adjacent cluster list of the search cluster;

[0037] Take one of the clusters To start the search cluster, use the neighborhood search method to traverse all the remaining classification clusters and obtain the adjacent clusters of all the remaining classification clusters List of adjacent clusters that coexist with the search cluster Where i is the serial number of the rock surface structural surface, j is the serial number of the cluster in each group of rock surface structural surfaces;

[0038] Step 2.2: Find three adjacent classification clusters, determine the vertices of the rock mass candidate blocks, and store them in the rock mass candidate block vertex list verticesList;

[0039] Find three adjacent classification clusters, and select one of them To start the search cluster, traverse the cluster The list of adjacent clusters of the search cluster Simultaneously search the adjacent cluster list The list of adjacent clusters for each cluster in If the adjacent cluster list and adjacent cluster lists contains the same adjacent clusters, such as Then it means For three adjacent clusters, the three adjacent classification clusters are intersected to obtain the intersection point, and the intersection point is the vertex of the rock mass candidate block; in the same way, the vertices of all rock mass candidate blocks are obtained and stored in the rock mass candidate block vertex list verticesList;

[0040] Step 2.3: Use the block growth algorithm (BGA) to identify candidate rock blocks, such as Figure 2-3 As shown, the following steps are included:

[0041] Step 2.3.1: Establish the initial space of the candidate rock mass block, calculate the opening direction of the initial space of the candidate rock mass block, and compare the angle between the opening direction and the normal direction with 90 degrees to determine whether the initial state of the candidate rock mass block meets the conditions; index the other vertices on the edge line where the vertex that meets the initial state of the candidate rock mass block is located, and execute step 2.3.2 for the next judgment;

[0042] Calculate the inclination, dip and normal direction of the original rock mass surface, and specify the normal direction V of the original rock mass surface n Towards the inside of the rock mass surface; the rock mass candidate block initially grows from a vertex, and the first vertex is formed by the intersection of three surfaces. The initial three surfaces can be used to determine a unique space, namely the initial space of the rock mass candidate block;

[0043] Calculate the opening direction V of the initial space of the rock mass candidate block i When V i With V n When the angle is greater than 90 degrees, it means that this vertex is not a valid starting point and does not meet the starting state of the candidate rock mass block, and the next intersection point is determined; when V i With V n When the angle is less than 90 degrees, it means that this vertex is a valid starting point and meets the starting state of the candidate rock mass block;

[0044] Add the vertices that meet the initial state of the rock candidate block to the vertex list verticesList of the rock candidate block, add the first cluster to the planesList of the rock candidate block surface, and index the other vertices on the edge line where the vertex that meets the initial state of the rock candidate block is located, and execute step 2.3.2 for the next judgment; the edge line is the intersection line of the two rock surface surfaces;

[0045] Step 2.3.2: Determine whether the vertex on the endpoint of the edge line is the endpoint of the edge line. If the vertex is on the endpoint of the edge line, execute step 2.3.3 for the next step of determination;

[0046] Each edge line contains multiple vertices. If the vertex is inside the edge line, determine the next vertex in the vertex list verticesList of the rock mass candidate block; if the vertex is at the endpoint of the edge line, execute step 2.3.3 to make the next judgment;

[0047] Step 2.3.3: Determine whether the vertex at the edge line endpoint is on the opening side of the rock mass candidate block. If so, add the vertex at the edge line endpoint to the rock mass candidate block vertex list verticesList;

[0048] If the vertex at the edge line endpoint is outside the rock mass candidate block, then the vertex at the edge line endpoint does not belong to the rock mass candidate block, and the next vertex is determined; if the vertex at the edge line endpoint is inside the rock mass candidate block, then the vertex at the edge line endpoint belongs to the candidate block, and the vertex at the edge line endpoint is added to the rock mass candidate block vertex list verticesList, and the other vertices on the edge line where the vertex is located are indexed, and step 2.3.4 is executed for the next step of determination;

[0049] Step 2.3.4: Determine whether the cluster contained in the newly added vertex already exists in the candidate block. If not, add the new cluster to the rock candidate block surface list planesList;

[0050] Each vertex contains three clusters. Check whether the three clusters corresponding to the vertex at the edge endpoint already exist in the rock candidate block list. If not, add the new cluster to the rock candidate block surface list planesList, and the number of clusters k becomes k+1. Repeat this process, traverse all clusters until all candidate blocks are indexed, then execute step 2.3.5 to proceed to the next judgment. The rock candidate block list includes the rock candidate block surface list planesList and the rock candidate block vertex list verticesList.

[0051] Step 2.3.5: Determine whether there are duplicate blocks in the candidate rock mass blocks. If so, delete the duplicate blocks. Otherwise, go to step 3.

[0052] At this time, the rock candidate block consists of the vertex list verticesList and the surface list planesList of the rock candidate block. The number of vertices and the order of vertices in the vertex list verticesList of the rock candidate block are judged. If the number of vertices is the same but the order is different, the rock candidate block is retained, and the number of rock candidate blocks m is increased to m+1, until all rock candidate blocks are judged.

[0053] Step 3: Determine the spatial topological relationship between the candidate rock mass block and the surrounding candidate rock mass blocks or the rock mass surface structure, thereby determining the dangerous block, that is, the dangerous and prone to sliding 3D point cloud block;

[0054] Index the lower surface of each candidate block; if the lower surface does not intersect with any position of the rock surface, mark this rock candidate block as a dangerous block; and so on, judge all rock candidate blocks, mark all dangerous blocks, and all marked dangerous blocks are called dangerous and easy-to-slide 3D point cloud blocks.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. A method for automatically identifying dangerous and prone-to-sliding rock mass in a three-dimensional point cloud, characterized by: The following steps are involved: Step 1: Use the three-dimensional Gaussian mixture model algorithm to automatically identify the rock surface structural planes, and obtain each classification cluster in each group of rock surface structural planes, and store each classification cluster in all rock surface structural planes in a specified indexing manner; Step 2: Find the adjacent clusters of each classification cluster in all rock surface structural surfaces, determine the vertices of the rock candidate blocks, and automatically identify the rock candidate blocks using the block growing algorithm; Step 2.1: Find the adjacent clusters of each classification cluster in all rock surface structural surfaces and store them in the adjacent cluster list of the search cluster; Step 2.2: Find three adjacent classification clusters, determine the vertices of the rock mass candidate blocks, and store them in the rock mass candidate block vertex list verticesList; Step 2.3: Identify candidate rock masses using a block growing algorithm; Step 2.3.1: Establish the initial space of the candidate rock mass block, calculate the opening direction of the initial space of the candidate rock mass block, and compare the angle between the opening direction and the normal direction with 90 degrees to determine whether the initial state of the candidate rock mass block meets the conditions; index the other vertices on the edge line where the vertex that meets the initial state of the candidate rock mass block is located, and execute step 2.3.2 for the next judgment; Calculate the inclination, dip and normal direction of the original rock mass surface, and specify the normal direction V of the original rock mass surface n Towards the inside of the rock mass surface; the rock mass candidate block initially grows from a vertex, and the first vertex is formed by the intersection of three rock mass surfaces. The initial three surfaces are used to determine a unique space, namely the initial space of the rock mass candidate block; Calculate the opening direction V of the initial space of the rock mass candidate block i When V i With V n If the angle is greater than 90 degrees, it means that this vertex is not a valid starting point and does not meet the starting state of the rock mass candidate block, and the next intersection point is determined; When V i With V n When the angle is less than 90 degrees, it means that this vertex is a valid starting point and meets the starting state of the candidate rock mass block; Add the vertices that meet the initial state of the rock candidate block to the vertex list verticesList of the rock candidate block, add the first cluster to the planesList of the rock candidate block surface, and index the other vertices on the edge line where the vertex that meets the initial state of the rock candidate block is located, and execute step 2.3.2 for the next judgment; the edge line is the intersection line of the two rock surface surfaces; Step 2.3.2: Determine whether the vertex of the edge line is on the endpoint of the edge line. If the vertex is on the endpoint of the edge line, execute step 2.3.3 to make the next judgment; Each edge line contains multiple vertices. If the vertex is inside the edge line, determine the next vertex in the vertex list verticesList of the rock mass candidate block; if the vertex is at the endpoint of the edge line, execute step 2.3.3 to make the next judgment; Step 2.3.3: Determine whether the vertex at the end point of the edge line is on the opening side of the candidate rock mass block. If so, add the vertex at the end point of the edge line to the vertex list of the candidate rock mass block verticesList; Step 2.3.4: Determine whether the cluster contained in the newly added vertex already exists in the candidate block. If not, add the new cluster to the rock candidate block surface list planesList; Step 2.3.5: Determine whether there are duplicate blocks in the candidate rock mass blocks. If so, delete the duplicate blocks. Otherwise, go to step 3. Step 3: Determine the spatial topological relationship between the candidate rock mass block and the surrounding candidate rock mass blocks or the rock mass surface structure, so as to determine the dangerous block, that is, the dangerous and prone to sliding 3D point cloud block.

2. The method for automatically identifying dangerous and prone-to-slip 3D point cloud blocks of rock mass according to claim 1, characterized in that: The specific method of step 2.1 is: Take one of the clusters To start the search cluster, use the neighborhood search method to traverse all the remaining classification clusters and obtain the adjacent clusters of all the remaining classification clusters List of adjacent clusters that coexist with the search cluster Where i is the serial number of the rock surface structural surface, and j is the serial number of the cluster in each group of rock surface structural surfaces.

3. The method for automatically identifying dangerous and prone-to-slip 3D point cloud blocks of rock mass according to claim 2, characterized in that: The specific method of step 2.2 is: Find three adjacent classification clusters, and select one of them To start the search cluster, traverse the cluster The list of adjacent clusters of the search cluster Simultaneously search the adjacent cluster list The list of neighboring clusters for each cluster in The intersection of three adjacent classification clusters is used as the vertex of the candidate rock mass block; In the same way, the vertices of all rock candidate blocks are obtained and stored in the rock candidate block vertex list verticesList.

4. The method for automatically identifying dangerous and prone-to-sliding three-dimensional point cloud blocks of rock mass according to claim 3, characterized in that: The specific method of step 2.3.3 is: If the vertex at the edge line endpoint is outside the rock mass candidate block, then the vertex at the edge line endpoint does not belong to the rock mass candidate block, and the next vertex is determined; If the vertex at the edge endpoint is inside the rock candidate block, then the vertex at the edge endpoint belongs to the candidate block. The vertex at the edge endpoint is added to the rock candidate block vertex list verticesList, and the other vertices on the edge where the vertex is located are indexed. Execute step 2.3.4 for the next judgment.

5. The method for automatically identifying dangerous and prone-to-sliding three-dimensional point cloud blocks of rock mass according to claim 4, characterized in that: The specific method of step 2.3.4 is: Each vertex contains three clusters. Check whether the three clusters corresponding to the vertex at the edge endpoint already exist in the rock candidate block list. If not, add the new cluster to the rock candidate block surface list planesList, and the number of clusters k becomes k+1. Similarly, all clusters are traversed until all candidate blocks are indexed, and then step 2.3.5 is executed to make the next judgment; the rock candidate block list includes a rock candidate block surface list planesList and a rock candidate block vertex list verticesList.

6. The method for automatically identifying dangerous and prone-to-sliding three-dimensional point cloud blocks of rock mass according to claim 5, characterized in that: The specific method of step 2.3.5 is: At this time, the rock candidate block consists of the rock candidate block vertex list verticesList and the rock candidate block surface list planesList. The number of vertices and the order of vertices in the rock candidate block vertex list verticesList are judged. When the number of vertices is the same but the order is different, the rock candidate block is retained, and the number of rock candidate blocks m becomes m+1, until all rock candidate blocks are judged.

7. The method for automatically identifying dangerous and prone-to-slip 3D point cloud blocks of rock mass according to claim 6, characterized in that: The specific method of step 3 is: Index the lower surface of each candidate block; if the lower surface of a candidate block does not intersect with any position on the rock surface, mark this rock candidate block as a dangerous block; and so on, judge all rock candidate blocks, mark all dangerous blocks, and all marked dangerous blocks are called dangerous and prone to sliding 3D point cloud blocks.

Images