CN-121982696-A - Method, system, equipment and medium for identifying rock mass structural surface of hydraulic tunnel of pumped storage power station

CN121982696ACN 121982696 ACN121982696 ACN 121982696ACN-121982696-A

Abstract

The invention is suitable for the field of rock mass structural planes, and discloses a method, a system, equipment and a medium for identifying the rock mass structural plane of a hydraulic tunnel of a pumped storage power station, wherein the method comprises the steps of obtaining an excavation region image and carrying out three-dimensional reconstruction to obtain a three-dimensional point cloud model; based on the three-dimensional point cloud model, a voxel cloud connectivity segmentation algorithm is adopted to segment the three-dimensional point cloud model to obtain a plurality of super-voxel blocks, the super-voxel blocks are used as micro planes, a principal component analysis algorithm is adopted to obtain a plane normal vector of each super-voxel block, and based on the plane normal vector, adjacent super-voxel blocks with similar normal vectors are aggregated through a region growing algorithm to obtain a structural surface. According to the invention, the three-dimensional point cloud of the excavated area is divided into the super-voxel blocks, so that the number of seed points required to be iterated by a subsequent area growth algorithm is reduced, the processing efficiency is improved, and high accuracy of structural surface measurement is maintained due to the aggregation point of the super-voxel blocks, so that the processing of the point cloud of the excavated area of the hydraulic tunnel in a very large scale and the structural surface identification are realized.

Inventors

XIA YUDONG
Huang Jicun
LI JING
YOU JIAXING
WANG BO
Zhan Zishi
FU ZHAOKAI
Chang Zheyu

Assignees

中国三峡建工(集团)有限公司

Dates

Publication Date: 20260505
Application Date: 20251127

Claims (10)

1. The method for identifying the rock mass structural surface of the hydraulic tunnel of the pumped storage power station is characterized by comprising the following steps of: Acquiring an excavation area image and carrying out three-dimensional reconstruction to obtain a three-dimensional point cloud model; based on the three-dimensional point cloud model, dividing by adopting a voxel cloud connectivity dividing algorithm to obtain a plurality of super-voxel blocks; The super-voxel blocks are used as micro planes, and a principal component analysis algorithm is adopted to obtain the plane normal vector of each super-voxel block; And based on the plane normal vector, the adjacent super-voxel blocks with similar normal vectors are aggregated through a region growing algorithm to obtain a structural surface.
2. The method for identifying the rock mass structural surface of a hydraulic tunnel of a pumped storage power station according to claim 1, wherein the steps of obtaining an image of an excavated area and performing three-dimensional reconstruction to obtain a three-dimensional point cloud model comprise the following steps: Acquiring a series of images of the excavation area of the hydraulic tunnel through multi-angle shooting, and setting the pixel duty ratio of the same area of the images with the same area; based on images shot by multiple views, three-dimensional reconstruction is performed by adopting digital photogrammetry software, and a three-dimensional point cloud model of the excavated area containing space coordinates is obtained.
3. The method for identifying the rock mass structural surface of the hydraulic tunnel of the pumped storage power station according to claim 2, wherein the method for identifying the rock mass structural surface of the hydraulic tunnel of the pumped storage power station is characterized by adopting a voxel cloud connectivity segmentation algorithm to segment the rock mass structural surface based on the three-dimensional point cloud model to obtain a plurality of super-voxel blocks, and comprises the following steps: the method comprises the steps of dividing by adopting a voxel cloud connectivity dividing algorithm by taking space coordinates of a three-dimensional point cloud model as input to obtain a plurality of super-voxel blocks; The structure of the obtained super-voxel block of the three-dimensional point cloud model comprises an index array of the super-voxel block and an index array of the point cloud to which each super-voxel block belongs.
4. A pumped storage power station hydraulic tunnel rock mass structural face identification method according to claim 3, wherein the method for obtaining the plane normal vector of each super voxel block by taking the super voxel block as a micro plane and adopting a principal component analysis algorithm comprises the following steps: And taking the super-voxel block as a micro-plane, calculating a characteristic value and a characteristic vector by adopting a principal component analysis algorithm according to the space coordinates of the point cloud to which the super-voxel block belongs, and taking the characteristic vector corresponding to the minimum characteristic value as a plane normal vector of the super-voxel block.
5. The method for identifying the rock mass structural plane of the hydraulic tunnel of the pumped storage power station according to claim 4, wherein the step of obtaining the plane normal vector of each super-voxel block by adopting a principal component analysis algorithm comprises the following steps: Calculating the average coordinates of all points of the super-voxel block as a centroid; Subtracting the mass centers from each point to obtain a decentered point set; Constructing a covariance matrix by adopting a decentralised point set; and carrying out feature decomposition on the covariance matrix to obtain three feature values and corresponding feature vectors, and sequencing the three feature values according to the size, wherein the feature vector corresponding to the minimum feature value is used as a plane normal vector of the super-voxel block.
6. The method for identifying the rock mass structural surface of the hydraulic tunnel of the pumped storage power station according to claim 5, wherein the method for obtaining the structural surface by aggregating adjacent super-voxel blocks with similar normal vectors through a region growing algorithm based on the plane normal vectors comprises the following steps: Selecting an ultra-voxel block as a seed point of an initial growth area by random non-repeated sampling; Based on all point clouds in the current growing area, recalculating the current plane normal vector of the current growing area through a principal component analysis algorithm; Judging through the super voxel blocks in the neighborhood of the current growing area and the plane normal vector of the current growing area, and converging the neighborhood super voxel blocks with the normal vector close to the current growing area; And taking the newly polymerized super voxel block as a new seed point, and iterating until the current growing area cannot be expanded, so as to obtain the finally generated growing area as the identified rock mass structural plane.
7. The method for identifying the rock mass structural surface of the hydraulic tunnel of the pumped storage power station according to claim 6, wherein the judging of the plane normal vector between the super voxel block passing through the neighborhood of the current growing area and the current growing area comprises the following steps: calculating cosine similarity between the plane normal vector of the neighborhood super-voxel block and the plane normal vector of the current growing area; and if the cosine similarity is smaller than a preset angle threshold, aggregating the neighborhood super-voxel blocks to the current growing area.
8. A pumped storage power station hydraulic tunnel rock mass structural face identification system employing the method of any one of claims 1-7, comprising: the three-dimensional reconstruction module is used for acquiring an image of the excavated area and carrying out three-dimensional reconstruction to obtain a three-dimensional point cloud model; the segmentation module is used for carrying out segmentation by adopting a voxel cloud connectivity segmentation algorithm based on the three-dimensional point cloud model to obtain a plurality of super-voxel blocks; the calculation module is used for obtaining a plane normal vector of each super-voxel block by taking the super-voxel block as a micro-plane and adopting a principal component analysis algorithm; and the aggregation module is used for aggregating adjacent super-voxel blocks with similar normal vectors through a region growing algorithm based on the plane normal vectors to obtain a structural surface.
9. An electronic device, comprising: A memory and a processor; The memory is configured to store computer executable instructions that, when executed by the processor, implement the steps of the pumped storage power station hydraulic tunnel rock face identification method of any one of claims 1 to 7.
10. A computer readable storage medium, comprising computer executable instructions stored thereon, which when executed by a processor, implement the steps of the pumped storage power station hydraulic tunnel rock face identification method of any one of claims 1 to 7.

Description

Method, system, equipment and medium for identifying rock mass structural surface of hydraulic tunnel of pumped storage power station Technical Field The invention relates to the field of rock mass structural planes, in particular to a method, a system, equipment and a medium for identifying a rock mass structural plane of a hydraulic tunnel of a pumped storage power station. Background The structural plane is a generic term for various geological interfaces of the cut rock mass. Under the influence of structural stress change and long-term geological action, a plurality of groups of dominant structural surfaces are usually developed in the rock mass of the hydraulic tunnel, and the spatial distribution of the structural surfaces is quite different, so that the structural surfaces of the rock mass are the key factors for determining the deformation and the damage of the rock mass of the hydraulic tunnel. Therefore, during hydraulic tunnel construction, the geological investigation of the rock mass is very important, so that the rock mass related information can be obtained, the engineering stability can be evaluated, and the basis can be provided for the subsequent support design. The structural surface measuring tools commonly used at present are devices such as a traditional geological compass and an inclinometer, and the defects of the devices are obvious. In practical use, the traditional measurement techniques have the problems of low efficiency, poor safety and special dependence on operator experience, seriously influence the speed and accuracy of geological data acquisition, and cannot meet the requirements of modern tunnel intelligent construction on high-precision and real-time geological information. Specifically, the technical defects are mainly three aspects, namely, the time consumption of measurement and easy conflict with the construction process. The rock mass exposure time is strictly controlled, the rock mass cannot be supported in time after excavation, the slow progress is dragged, the safety is influenced, the method is limited by the reachable range of excavated rock mass, the technical level and experience of measuring staff are different, which structural surfaces are selected to be used as samples, data are collected, subjectivity and locality are strong, the data are not representative enough, the reliability of statistical analysis results is influenced, and if surrounding rock stability is poor, measuring staff can get close to the temporary surface to work, the safety risks of rock mass spalling, water burst and mud burst and even local collapse are met, which not only threatens personal safety, but also further influences the integrity and feasibility of data collection. There is therefore a need for a rock mass face identification method that is efficient, safe and not highly dependent on operator experience. Disclosure of Invention The present invention has been made in view of the above-mentioned existing problems of inefficiency, poor safety, and high dependency on operator experience. Therefore, the invention provides a method, a system, equipment and a medium for identifying the rock mass structural surface of a hydraulic tunnel of a pumped storage power station, which solve the outstanding problems of low efficiency, poor safety, high dependence on the experience of an operator and the like in the prior art. In order to solve the technical problems, the invention provides the following technical scheme: in a first aspect, the invention provides a method for identifying a rock mass structural surface of a hydraulic tunnel of a pumped storage power station, which comprises the following steps: Acquiring an excavation area image and carrying out three-dimensional reconstruction to obtain a three-dimensional point cloud model; based on the three-dimensional point cloud model, dividing by adopting a voxel cloud connectivity dividing algorithm to obtain a plurality of super-voxel blocks; The super-voxel blocks are used as micro planes, and a principal component analysis algorithm is adopted to obtain the plane normal vector of each super-voxel block; And based on the plane normal vector, the adjacent super-voxel blocks with similar normal vectors are aggregated through a region growing algorithm to obtain a structural surface. The invention relates to a method for identifying the rock mass structural surface of a hydraulic tunnel of a pumped storage power station, which comprises the following steps of obtaining an image of an excavated area and carrying out three-dimensional reconstruction to obtain a three-dimensional point cloud model, wherein the method comprises the following steps: Acquiring a series of images of the excavation area of the hydraulic tunnel through multi-angle shooting, and setting the pixel duty ratio of the same area of the images with the same area; based on images shot by multiple views, three-dimensional reconstruction is performed by adopting digit