CN-122023427-A - Unmanned aerial vehicle perception-based road subsidence risk assessment method and system

CN122023427ACN 122023427 ACN122023427 ACN 122023427ACN-122023427-A

Abstract

A road subsidence risk assessment method and system based on unmanned aerial vehicle perception belong to the technical field of road subsidence risk assessment. The method comprises the steps of installing a structural light camera on an unmanned plane, collecting ground road point cloud data to obtain road surface coordinates, fitting a reference curved surface, calculating road subsidence amount by comparing actual point cloud data, conducting grid subsidence statistics, conducting main subsidence direction analysis on each grid to obtain resultant force intensity of each grid subsidence direction, calculating structural strain energy density and tensor singular degree, and building a road subsidence comprehensive risk model based on average subsidence value, subsidence fluctuation non-uniformity, subsidence direction resultant force intensity, structural strain energy density and tensor singular degree of each grid. The invention improves the accuracy of monitoring the road subsidence and is beneficial to the rationality of maintenance decision.

Inventors

MENG ANXIN
CHENG GONG
LIU XING
WANG WENJING
ZHAO HAIYUN
FENG JUNHUA

Assignees

深城交科技集团股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260415

Claims (9)

1. The road subsidence risk assessment method based on unmanned aerial vehicle perception is characterized by comprising the following steps of: s1, installing a structured light camera on an unmanned aerial vehicle, and collecting ground road point cloud data to obtain road surface coordinates; s2, fitting a reference curved surface as a zero reference surface for identifying road subsidence; s3, calculating the road subsidence amount by comparing the actual point cloud data based on the reference curved surface obtained by fitting in the step S2; s4, carrying out grid subsidence statistics on the road subsidence obtained in the step S3; s5, analyzing the main sinking direction of the road for each grid obtained in the step S4 to obtain the resultant force strength of the sinking direction of each grid; s6, calculating the structural strain energy density and tensor singular degree of each grid obtained in the step S4; And S7, establishing a road subsidence comprehensive risk model based on the average subsidence value, subsidence fluctuation unevenness, subsidence direction resultant force strength, structural strain energy density and tensor singular degree of each grid.
2. The method for evaluating the risk of road subsidence based on unmanned aerial vehicle perception according to claim 1, wherein the road surface coordinates P obtained in step S1 are: ; Wherein, the The three-dimensional coordinates of the ith point cloud are obtained by the structured light camera, and N is the total number of the point clouds.
3. The road subsidence risk assessment method based on unmanned aerial vehicle perception according to claim 2, wherein the specific implementation method of step S2 comprises the following steps: s2.1, setting the total number of the reference points used for fitting the reference curved surface as K, and calculating the distance from the ground coordinate to the jth reference point Establishing the elevation value of the ground coordinates (x, y) on the reference curved surface The expression is: ; ; Wherein, the Is the two-dimensional coordinates of the j-th reference point, Is a two-dimensional coordinate of the ground, 、、 Respectively a reference offset term, an x-direction slope term and a y-direction slope term, which are obtained by fitting and are simultaneously used for adjusting dimension; the radial base coefficient of the j-th reference point is obtained by fitting and is used for adjusting dimension; As a reference distance, determined empirically by an expert; s2.2, constructing a model error objective function E: ; Wherein, the Three-dimensional coordinates of a j-th reference point; Is a curved surface smoothing regulation factor, and is determined by expert experience; s (x j , y j ) is the reference surface obtained by fitting in the form of partial guide symbol Elevation value at; S2.3 based on And minimizing the model error objective function E to solve the optimal parameters And obtaining a stable reference curved surface.
4. The method for evaluating risk of road subsidence based on unmanned aerial vehicle perception according to claim 3, wherein the formula for calculating the road subsidence amount in step S3 is: ; Wherein, the The road subsidence amount for the i-th point cloud.
5. The method for estimating risk of road subsidence based on unmanned aerial vehicle perception according to claim 4, wherein the method for performing mesh subsidence statistics in step S4 is to calculate average subsidence value and subsidence fluctuation unevenness of each mesh, and the calculation formula is: ; ; Wherein, the An average amount of subsidence for the grid (m, n); a set of points covered by a mesh (m, n); Is the sink fluctuation unevenness of the grid (m, n).
6. The method for estimating risk of road subsidence based on unmanned aerial vehicle perception according to claim 5, wherein step S5 gives a resultant force vector caused by subsidence to an average value obtained by weighting the point clouds in each grid according to the subsidence values, represents the direction of overall tendency of subsidence in each grid, and obtains the resultant force intensity of subsidence direction of grid (m, n) 。
7. The method for evaluating risk of road subsidence based on unmanned aerial vehicle perception according to claim 6, wherein the specific implementation method of step S6 comprises the following steps: S6.1, characterizing the strain energy density of the structure by utilizing the point cloud covariance tensor characteristics, wherein the expression is as follows: ; Wherein, the Structural strain energy density for grid (m, n); The spatial spreading scale of the point cloud of the grid (m, n) in the main direction; 、、 The strain energy quadratic term factor, the strain energy cubic term factor and the strain energy logarithmic term factor are respectively determined by expert experience and are simultaneously used for adjusting dimension; S6.2, considering that the road subsidence development direction comprises two dimensions of a vertical dimension and a plane, constructing tensor singular degree by utilizing point cloud covariance tensor characteristics, wherein the expression is as follows: ; Wherein, the Tensor singular values for grid (m, n); a spatial spread scale in the secondary direction for the point cloud of the grid (m, n); The spatial spread scale of the point cloud of the grid (m, n) in the minimum feature direction.
8. The method for evaluating risk of road subsidence based on unmanned aerial vehicle perception according to claim 7, wherein the expression of the comprehensive risk model of road subsidence in step S7 is: ; Wherein, the Is a comprehensive risk value of subsidence; is a fusion intensity factor, and is determined by expert experience; Is a subsidence amount reference value, and is determined by expert experience; is a strain energy density reference value and is determined by expert experience; Is a non-uniformity reference value, and is determined empirically by an expert; The singular degree reference value is tensor and is determined empirically by expert; Is a resultant force strength reference value of the sinking direction, and is determined by expert experience; ~ Respectively a subsidence weight coefficient, a strain energy weight coefficient, a deformation disturbance weight coefficient, a singular degree weight coefficient and a subsidence direction resultant force intensity coefficient, which are determined by expert experience.
9. A system based on an unmanned perceived roadway subsidence risk assessment method comprising a processor, a memory and a computer program stored in the memory and executable on the processor, the computer program when executed implementing the steps of an unmanned perceived roadway subsidence risk assessment method according to any one of claims 1-8.

Description

Unmanned aerial vehicle perception-based road subsidence risk assessment method and system Technical Field The invention belongs to the technical field of road subsidence risk assessment, and particularly relates to a road subsidence risk assessment method and system based on unmanned aerial vehicle perception. Background Road subsidence is one of the common problems in traffic infrastructures such as cities and highways, and refers to the phenomenon that collapse or slippage, camber and the like occur in a certain range on a road surface, and the problems of difficult running, water drainage blockage, accidents and the like of automobiles can be caused under extreme conditions. However, the conventional inspection means mainly depend on manpower and contact type displacement monitoring equipment, which is time-consuming and labor-consuming, and only a small part of road sections can be investigated, so that the whole condition of the road cannot be known timely and accurately. In recent years, unmanned aerial vehicles have been used for road surface structure detection because of their high viewing angle, non-contact, etc. The unmanned aerial vehicle carrying the high-precision structured light camera can obtain millimeter-level road point clouds in a short time, and the situation of micro-deformation of the road is represented by adopting a larger scale. Disclosure of Invention The invention aims to solve the problems of low accuracy, difficult direction identification and difficult risk quantification in road subsidence detection, and provides a road subsidence risk assessment method and system based on unmanned aerial vehicle perception. In order to achieve the above purpose, the present invention is realized by the following technical scheme: A road subsidence risk assessment method based on unmanned aerial vehicle perception comprises the following steps: s1, installing a structured light camera on an unmanned aerial vehicle, and collecting ground road point cloud data to obtain road surface coordinates; s2, fitting a reference curved surface as a zero reference surface for identifying road subsidence; s3, calculating the road subsidence amount by comparing the actual point cloud data based on the reference curved surface obtained by fitting in the step S2; s4, carrying out grid subsidence statistics on the road subsidence obtained in the step S3; s5, analyzing the main sinking direction of the road for each grid obtained in the step S4 to obtain the resultant force strength of the sinking direction of each grid; s6, calculating the structural strain energy density and tensor singular degree of each grid obtained in the step S4; And S7, establishing a road subsidence comprehensive risk model based on the average subsidence value, subsidence fluctuation unevenness, subsidence direction resultant force strength, structural strain energy density and tensor singular degree of each grid. Further, the road surface coordinates P obtained in step S1 are: Wherein, the The three-dimensional coordinates of the ith point cloud are obtained by the structured light camera, and N is the total number of the point clouds. Further, the specific implementation method of the step S2 includes the following steps: s2.1, setting the total number of the reference points used for fitting the reference curved surface as K, and calculating the distance from the ground coordinate to the jth reference point Establishing the elevation value of the ground coordinates (x, y) on the reference curved surfaceThe expression is: Wherein, the Is the two-dimensional coordinates of the j-th reference point,Is a two-dimensional coordinate of the ground,、、Respectively a reference offset term, an x-direction slope term and a y-direction slope term, which are obtained by fitting and are simultaneously used for adjusting dimension; the radial base coefficient of the j-th reference point is obtained by fitting and is used for adjusting dimension; As a reference distance, determined empirically by an expert; s2.2, constructing a model error objective function E: Wherein, the Three-dimensional coordinates of a j-th reference point; Is a curved surface smoothing regulation factor, and is determined by expert experience; s (x j, yj) is the reference surface obtained by fitting in the form of partial guide symbol Elevation value at; S2.3 based on And minimizing the model error objective function E to solve the optimal parametersAnd obtaining a stable reference curved surface. Further, the formula for calculating the road subsidence amount in step S3 is as follows: Wherein, the The road subsidence amount for the i-th point cloud. Further, the method for performing the grid settlement statistics in step S4 is to calculate the average settlement value and settlement fluctuation unevenness of each grid, and the calculation formula is as follows: Wherein, the An average amount of subsidence for the grid (m, n); a set of points covered by a mesh (m,