CN-121982196-A - Underground engineering structure full-space deformation analysis method based on multi-depth camera array

CN121982196ACN 121982196 ACN121982196 ACN 121982196ACN-121982196-A

Abstract

The application discloses a multi-depth camera array-based full-space deformation analysis method for an underground engineering structure, and relates to the field of space deformation; the method comprises the steps of performing internal reference calibration on depth cameras, converting aligned depth images of the depth cameras into three-dimensional color point cloud data of the depth cameras, performing external reference calibration on each depth camera to obtain an optimized transformation matrix from each slave camera to a master camera, transforming the three-dimensional color point cloud data of each slave camera to a coordinate system of the master camera by adopting the optimized transformation matrix to obtain a three-dimensional model of an underground engineering structure, determining corresponding point pairs in the three-dimensional model obtained in different periods, and calculating Euclidean distances between the corresponding point pairs to realize the full-space deformation analysis of the underground engineering structure. The application realizes the deformation monitoring of the underground engineering with low cost and high precision.

Inventors

MA SHUQI
XU YUANZHEN
ZHANG KEXUE
ZHANG WEIGUANG
WANG HONGZHI
LI YANHUI
CHEN JIAZHENG
Yao Xiangchen
ZHANG ZHAOYUAN

Assignees

安徽理工大学

Dates

Publication Date: 20260505
Application Date: 20251229

Claims (10)

1. An underground engineering structure full-space deformation analysis method based on a multi-depth camera array, which is characterized by comprising the following steps: Collecting depth images and color images of an underground engineering structure by adopting a multi-depth camera array, and aligning the depth images with the color images to generate aligned depth images of each depth camera, wherein the multi-depth camera array comprises a master camera and a plurality of slave cameras; Performing internal reference calibration on any depth camera, and converting an aligned depth image of the depth camera into three-dimensional color point cloud data under the self coordinate system of the depth camera based on the depth camera with calibrated internal reference; performing external parameter calibration on each depth camera to obtain an optimized transformation matrix from each slave camera to the master camera; Transforming the three-dimensional color point cloud data under the own coordinate system of each slave camera into the coordinate system of the master camera by adopting an optimized transformation matrix from each slave camera to the master camera, and carrying out point cloud fusion to obtain a three-dimensional model of the underground engineering structure; And determining corresponding point pairs in the three-dimensional model acquired in different periods based on normal vector information of the point cloud data, and calculating Euclidean distances between the corresponding points to realize the full-space deformation analysis of the underground engineering structure.
2. The method for analyzing the deformation of the underground engineering structure in the whole space based on the multi-depth camera array according to claim 1, wherein the method for calibrating the external parameters of each depth camera to obtain the optimized transformation matrix from each slave camera to the master camera specifically comprises the following steps: Performing external parameter calibration on each depth camera through a common view calibration plate to obtain an initial transformation matrix of each slave camera relative to the master camera; And optimizing an initial transformation matrix of each slave camera relative to the master camera based on the geometric features and the color features of the three-dimensional color point cloud data of the actual scene acquired by each depth camera to obtain an optimized transformation matrix from each slave camera to the master camera.
3. The method for analyzing the full-space deformation of the underground engineering structure based on the multi-depth camera array according to claim 2, wherein the method for calibrating the external parameters of each depth camera through the common view calibration plate to obtain the initial transformation matrix of each slave camera relative to the master camera specifically comprises the following steps: Aiming at any slave camera, the slave camera and the master camera are adopted to respectively acquire color images of the calibration plate in a common area; Based on the color image of the calibration plate, the internal reference of the slave camera, the internal reference of the master camera and the dimensional parameters of the calibration plate, obtaining the relative pose of the calibration plate and the slave camera and the relative pose of the calibration plate and the master camera by adopting an OpenCV function and a PnP algorithm; And obtaining an initial transformation matrix of the slave camera relative to the master camera based on the relative pose of the calibration plate and the slave camera and the relative pose of the calibration plate and the master camera.
4. The method for analyzing the deformation of the underground engineering structure in the whole space based on the multi-depth camera array according to claim 2, wherein the method is characterized in that the method adopts the three-dimensional color point cloud data of the actual scene collected by each depth camera, optimizes the initial transformation matrix of each slave camera relative to the master camera based on the geometric features and the color features of the three-dimensional color point cloud data of the actual scene to obtain the optimized transformation matrix from each slave camera to the master camera, and specifically comprises the following steps: for any slave camera, transforming the actual scene three-dimensional color point cloud data under the self coordinate system acquired by the slave camera into the coordinate system of the master camera through an initial transformation matrix of the slave camera relative to the master camera; Determining a flattening coefficient of the point cloud and a color change coefficient of the point cloud aiming at any point cloud in the actual scene three-dimensional color point cloud data acquired by the slave camera and the actual scene three-dimensional color point cloud data acquired by the master camera under a coordinate system of the master camera; Screening the actual scene three-dimensional color point cloud data acquired by the slave camera and the actual scene three-dimensional color point cloud data acquired by the master camera according to a set threshold value and the flattening coefficient of each point cloud to obtain the characteristic points of the slave camera and the characteristic points of the master camera; Aiming at any one of the characteristic points of the slave camera and the characteristic points of the master camera, obtaining a descriptor of the characteristic point according to an FPFH descriptor of the characteristic point and a color change coefficient of the characteristic point; Calculating the similarity between the descriptors of the characteristic points of the slave camera and the descriptors of the characteristic points of the master camera, and determining a plurality of matching point pairs according to the similarity; Solving pose relation by adopting SVD according to the space coordinates of any matching point pair to obtain a rotation matrix and a translation vector of the matching point pair; determining the root mean square error of the matching point pair according to the rotation matrix and the translation vector of the matching point pair; And determining the minimum root mean square error according to the root mean square errors of all the matching point pairs, and determining an optimized transformation matrix from the slave camera to the master camera according to the rotation matrix corresponding to the minimum root mean square error.
5. The method for analyzing the full-space deformation of the underground engineering structure based on the multi-depth camera array according to claim 4, wherein determining the flatness coefficient of the point cloud and the color change coefficient of the point cloud specifically comprises: Determining k nearest neighbors of the point cloud to construct a neighborhood point set of the point cloud; Calculating the mass center of a neighborhood point set, determining a first covariance matrix based on the mass center and k nearest neighbors, performing feature decomposition on the first covariance matrix to obtain a plurality of feature values, and determining a flattening coefficient of the point cloud according to the plurality of feature values, wherein the flattening coefficient is obtained according to the ratio of the sum of the plurality of feature values to the minimum feature value in the plurality of feature values; And constructing a second covariance matrix according to the color values of k nearest neighbors, carrying out feature decomposition on the second covariance matrix to obtain a plurality of feature values, and determining the color change coefficient of the point cloud according to the plurality of feature values.
6. The method for analyzing the full space deformation of the underground engineering structure based on the multi-depth camera array according to claim 5, the method is characterized in that the first covariance matrix is determined by adopting the following formula: ; ; ; Wherein, the Is the centroid of the set of neighborhood points, For the number of nearest neighbors, For the index of the nearest neighbor point, Is the first The number of the nearest neighbors to each other, As a variable Sum variable The covariance between the two is calculated by the method, 、、 For the coordinate component of the i nearest neighbor, 、 And Is the component of the coordinates of the centroid, For the first covariance matrix, As a variable Sum variable The covariance between the two is calculated by the method, As a variable Sum variable The covariance between the two is calculated by the method, As a variable Sum variable The covariance between the two is calculated by the method, As a variable Sum variable The covariance between the two is calculated by the method, As a variable Sum variable Covariance between.
7. The method for analyzing the total spatial deformation of a subsurface engineering structure based on a multi-depth camera array according to claim 5, wherein the following formula is used to calculate the slave camera's first Descriptors of feature points and the first of the main cameras Similarity of descriptors of individual feature points: ; Wherein, the The first of the slave cameras Descriptors of feature points and the first of the main cameras Similarity of descriptors of the individual feature points, For the dimension of the FPFH descriptor, Is the first The first of the FPFH descriptors of the feature points The dimensions of the dimensions, Is the first The first of the FPFH descriptors of the feature points The dimensions of the dimensions, Is the first The color change coefficients of the individual feature points, Is the first The color change coefficients of the individual feature points, As a result of the first weighting factor, Is a second weighting factor.
8. The method for analyzing the full-space deformation of the underground engineering structure based on the multi-depth camera array according to claim 1, wherein the different periods comprise a first period and a second period, point cloud data corresponding to the three-dimensional model acquired in the first period is used as source point cloud data, point cloud data corresponding to the three-dimensional model acquired in the second period is used as target point cloud data, corresponding point pairs in the three-dimensional model acquired in the different periods are determined based on normal vector information of the point cloud data, and full-space deformation analysis of the underground engineering structure is realized by calculating Euclidean distances between the corresponding point pairs, and the method specifically comprises the following steps: Downsampling the source point cloud data to obtain downsampled point cloud data; For any point in the downsampled point cloud data, determining a normal vector of the point, and establishing a spatial linear equation of the point based on the normal vector of the point and the three-dimensional coordinates of the point: Wherein, the method comprises the steps of, Is the three-dimensional coordinates of the corresponding point on the spatial straight line, For the three-dimensional coordinates of the point, As a normal vector to the point in question, Is an intermediate parameter; Calculating the distance from each point in the target point cloud data to the space straight line of the point, and screening the target point cloud data according to the distance and the set quantity to obtain a plurality of screened point clouds; establishing a space plane equation according to the screened multiple point clouds, obtaining intermediate parameters according to a space straight line equation of the points and the space plane equation, and determining three-dimensional coordinates of corresponding points on a space straight line according to the intermediate parameters; determining Euclidean distance between a corresponding point on a spatial straight line and the point according to the three-dimensional coordinates of the corresponding point and the three-dimensional coordinates of the point on the spatial straight line; and taking the Euclidean distance as a spatial deformation value of the point, and realizing the full-spatial deformation analysis of the underground engineering structure according to the spatial deformation value of each point in the down-sampling point cloud data.
9. The method for analyzing the total spatial deformation of the underground engineering structure based on the multi-depth camera array according to claim 8, wherein the following formula is adopted to calculate the first point in the cloud data of the target point Distance of a point to a spatial line of said point: ; ; ; ; ; Wherein, the Is the first in the target point cloud data The distance of a point from the spatial line of said point, As a first spatial straight line parameter, As a second spatial straight line parameter, As a third spatial straight line parameter, Is the first in the target point cloud data The three-dimensional coordinates of the individual points, For the index of points in the target point cloud data, Is a fourth spatial straight line parameter.
10. The method for analyzing the full space deformation of the underground engineering structure based on the multi-depth camera array according to claim 8, wherein the following formula is adopted to establish a space plane equation: ; ; ; ; ; Wherein, the As a parameter of the first spatial plane, As a parameter of the second spatial plane, As a parameter of the second spatial plane, As a parameter of the first spatial plane, To screen the three-dimensional coordinates of the first point cloud, For the three-dimensional coordinates of the second point cloud after screening, And the three-dimensional coordinates of the third point cloud after screening.

Description

Underground engineering structure full-space deformation analysis method based on multi-depth camera array Technical Field The application relates to the field of spatial deformation, in particular to a method for analyzing the total spatial deformation of an underground engineering structure based on a multi-depth camera array. Background Deformation monitoring of underground works is critical to ensure the safety and stability of the engineering structure. Conventional deformation monitoring methods, such as total stations, convergence meters, laser rangefinders, and the like, typically rely on analysis of in-situ cross-sectional data. Although these methods are still accurate, they are inefficient and difficult to meet the real-time monitoring requirements of large-scale subsurface projects. The three-dimensional laser scanning technology can acquire the structural point cloud data rapidly in real time, and has become an important tool for monitoring deformation of underground engineering. However, the tool still has many challenges in practical application, firstly, the point cloud density is high, the total data amount is large, so that direct analysis and post-processing have high requirements on software and hardware of a computer, and the application threshold of the computer with high performance and high storage is improved. However, the consumer-level depth camera can collect data in real time and has low requirements on illumination, so that the system is very suitable for data collection of underground engineering, and the multi-depth camera array can solve the limitation of small working range of single equipment. In deformation monitoring, two-dimensional section monitoring is a conventional means, section analysis can only reflect local deformation conditions, and overall deformation characteristics of the whole structure are difficult to comprehensively capture. The relative full-space deformation analysis can capture local deformation, reflect the cooperative deformation characteristics of the whole structure and provide comprehensive data support for engineering safety. Disclosure of Invention The application aims to provide a full-space deformation analysis method for an underground engineering structure based on a multi-depth camera array, which can accurately reconstruct the underground engineering structure based on the multi-depth camera array and perform full-space deformation analysis on the underground engineering structure under a high-precision three-dimensional model so as to realize low-cost and high-precision underground engineering deformation monitoring. In order to achieve the above object, the present application provides the following solutions: In a first aspect, the present application provides a method for analyzing the deformation of an underground engineering structure in full space based on a multi-depth camera array, comprising: Collecting depth images and color images of an underground engineering structure by adopting a multi-depth camera array, and aligning the depth images with the color images to generate aligned depth images of each depth camera, wherein the multi-depth camera array comprises a master camera and a plurality of slave cameras; Performing internal reference calibration on any depth camera, and converting an aligned depth image of the depth camera into three-dimensional color point cloud data under the self coordinate system of the depth camera based on the depth camera with calibrated internal reference; performing external parameter calibration on each depth camera to obtain an optimized transformation matrix from each slave camera to the master camera; Transforming the three-dimensional color point cloud data under the own coordinate system of each slave camera into the coordinate system of the master camera by adopting an optimized transformation matrix from each slave camera to the master camera, and carrying out point cloud fusion to obtain a three-dimensional model of the underground engineering structure; And determining corresponding point pairs in the three-dimensional model acquired in different periods based on normal vector information of the point cloud data, and calculating Euclidean distances between the corresponding points to realize the full-space deformation analysis of the underground engineering structure. According to the specific embodiment provided by the application, the application has the following technical effects: The application provides a full-space deformation analysis method of an underground engineering structure based on a multi-depth camera array, which is used for collecting depth images and color images of the underground engineering structure in multiple angles and multiple directions, laying a foundation for obtaining a clear and distinct three-dimensional model of the underground engineering structure later, and solving the problem of insufficient working range of a single camera by a multi-consumer-lev