CN-122023710-A - Cooperative modeling method for earthwork oblique photography of high-steep narrow terrain

CN122023710ACN 122023710 ACN122023710 ACN 122023710ACN-122023710-A

Abstract

The invention discloses a collaborative modeling method for earthwork oblique photography of high and steep narrow terrains, and belongs to the technical field of engineering mapping and three-dimensional digital modeling. According to the method, a ground-imitating flight route is designed based on a digital elevation model of a high and steep narrow terrain, an unmanned aerial vehicle carrying a laser radar and multiple lenses is controlled to synchronously acquire data, classified image control points are distributed, a digital elevation surface is generated after the laser radar point cloud is denoised, a three-dimensional grid model is generated by processing an oblique photographic image, the elevation is corrected by combining the two models, the texture is projected again, a collaborative live-action three-dimensional model is generated, and finally the earth-rock mass and a planned construction mechanical path are calculated based on the model. The method can be used for earthwork mapping, metering and construction scheduling of high and steep narrow terrains, and realizes cooperative promotion of modeling geometric accuracy and texture authenticity.

Inventors

CAI RONGSHENG
ZHOU GUANGYI
JIANG WENBO
SHANG YUQI
DONG DONGXUE
WANG LINAN
An Mengzhao
WANG JINGWEI
CHEN YANCHUN
YANG HAI
SUN JIAQI
WU LIANGFU

Assignees

中国水利水电第六工程局有限公司

Dates

Publication Date: 20260512
Application Date: 20251212

Claims (10)

1. The collaborative modeling method for the earthwork oblique photography of the high and steep narrow terrain is characterized by comprising the following steps of: S1, designing a ground-imitating flight route based on a digital elevation model of a high-steep narrow terrain, controlling an unmanned aerial vehicle carrying a laser radar and at least two oblique angle photographic lenses to fly along the route, and ensuring synchronous execution of laser radar scanning and multi-angle oblique photography through a time synchronization device; S2, arranging image control points in a region, namely arranging a first-stage control point network according to the density of 4-6 photographic baselines at heading intervals and 2 routes at side intervals along a ground-imitating flying route, and adding four-corner control points of an area network to serve as second-stage encryption points aiming at landform abrupt positions or key construction areas; S3, after coordinate conversion is carried out on the laser radar point cloud data, an iterative local plane fitting filtering method is adopted for denoising aiming at the high cliff wall point cloud noise, and the method comprises the steps of constructing a neighborhood for points in the point cloud, fitting a local plane based on the neighborhood points, calculating the distance from the points to the plane, and judging and eliminating noise points according to the distance statistical characteristics; S4, performing ground point classification and interpolation processing on the denoised point cloud to generate a digital elevation surface meeting the preset precision requirement; s5, performing space three encryption and dense matching on the oblique photographic image to generate an oblique three-dimensional grid model; S6, fusing the digital elevation surface and the inclined three-dimensional grid model, executing geometric constraint texture mapping operation, correcting the elevation of the vertex of the inclined three-dimensional grid model by taking the elevation value of the digital elevation surface as a reference, aligning the elevation value of the vertex with the elevation value of the digital elevation surface at the same position of the plane, and re-projecting textures according to the corrected grid geometry to generate a cooperative live-action three-dimensional model of high steep narrow topography; And S7, based on the collaborative live-action three-dimensional model, performing earth and stone quantity calculation by comparing model elevation changes in different periods, and performing construction machinery path planning by planning and displaying tracks in the model.
2. The collaborative modeling method for earthmoving oblique photography on high and steep narrow terrain according to claim 1, wherein step S1 comprises: S1a, designing a ground-imitation flight route based on a pre-acquired digital elevation model of high and steep narrow terrain, setting a heading overlapping degree range to be 75-85% and a side overlapping degree range to be 65-75%, calculating a photographing base line length B according to a formula B=GSD×H/f, and setting a photographing interval, wherein GSD is a preset ground sampling distance, f is a camera focal length, and H is a real-time navigational altitude relative to the terrain below; s1b, controlling an unmanned aerial vehicle carrying a laser radar and a multi-lens oblique photography camera to fly along a route, synchronously acquiring point cloud and image data, calibrating and compensating a space offset between the laser radar and the camera, controlling the space offset within a preset range, realizing synchronous acquisition by a time synchronization device, outputting a synchronization signal to the laser radar and the camera by the time synchronization device, recording time stamps of respective data acquisition, and storing the point cloud data and the image data with time stamp difference values smaller than a preset value in a correlated way.
3. The collaborative modeling method for high and steep narrow terrain earthwork oblique photography according to claim 1 or 2, wherein the high and steep narrow terrain digital elevation model is obtained by: a. Performing initial aerial survey on a region by adopting an unmanned plane laser radar, and collecting sparse point clouds; b. Denoising and classifying the sparse point cloud, and generating an initial digital elevation model through interpolation; c. And supplementing actual measurement control points in a steep region of the initial digital elevation model, and correcting the elevation value of the initial digital elevation model to obtain a final digital elevation model for the air route design.
4. The collaborative modeling method for earthmoving oblique photography of high and steep narrow terrain according to claim 1, wherein in step S3, the iterative local plane fitting filtering method comprises: S3a, constructing a self-adaptive neighborhood for each point P in the point cloud, calculating the local plane roughness sigma of the point P and K initial adjacent points, judging as a steep region and adopting a first neighborhood searching radius R s if sigma is larger than a preset steep judging threshold sigma t , otherwise adopting a second neighborhood searching radius R d , wherein the value range of R s ＜R d and K is 15-20, wherein the adjacent points are searched and acquired in a point cloud database through a spatial index algorithm, and gradually expanding the searching radius until the number requirement is met when the number of the searched points in the initial searching radius is insufficient; S3b, fitting a local plane to the point set in the self-adaptive neighborhood through a principal component analysis algorithm, and calculating the distance D from the point P to the plane; S3c, calculating the average value mu and the standard deviation delta of the distances from all points in the neighborhood to the plane; S3D, if D > mu+n multiplied by delta, marking the point P as a noise point, wherein the multiple factor n adopted in the steep region is the multiple factor adopted in the gentle region, and the value range of n is 2-3; And S3e, iteratively executing the steps S3a-S3d, and gradually shrinking the neighborhood search radius or adjusting the multiple factor n until the number of marked noise points is stable.
5. The collaborative modeling method for oblique photography of earthwork on high and steep narrow terrain according to claim 1, wherein in step S6, the correction of the vertex elevation of the oblique three-dimensional mesh model is achieved by: S6a, projecting a triangular patch T of the inclined three-dimensional grid model onto the digital elevation surface S, and obtaining an elevation point set { Z } of S in the coverage range of the triangular patch T; S6b, for the vertex V of the triangular patch T, obtaining a corresponding constraint elevation Z dem at a plane coordinate (x, y) of the triangular patch T through interpolation based on { Z }; S6c, calculating elevation residual error delta z=z dem -z mesh of original elevation z mesh and original elevation residual error of original elevation residual error z dem of the vertex V, and if the absolute value of delta z is larger than a preset tolerance threshold tau, executing correction, wherein tau is 0.05m; And S6d, in correction, determining a smoothing weight factor lambda based on the geometric position of the vertex V in the triangular patch T, so that the value of lambda increases as the vertex V approaches the center of the triangular patch T and decreases as the vertex V approaches the edge of the triangular patch T, and setting the new elevation z set of the vertex V as z set = z mesh +lambda multiplied by delta z.
6. The collaborative modeling method for the earthwork oblique photography of the steep and narrow terrain according to claim 1, wherein the meeting of the preset precision requirement in the step S4 means that the Gao Chengzhong error of the digital elevation surface is less than or equal to 0.10m; and (3) the error in the plane precision of the collaborative live-action three-dimensional model generated in the step S6 is less than or equal to 0.05m, and the error in the elevation precision is less than or equal to 0.10m.
7. The collaborative modeling method for earthwork oblique photography of high steep and narrow terrains according to claim 1, wherein the re-projecting the texture in step S6 comprises re-calculating the normal vector of each triangular patch according to the corrected triangular patch vertex coordinates, and performing orthographic correction and texture fusion on the original oblique image according to the new normal vector and camera parameters.
8. The collaborative modeling method for the earthwork oblique photography of high steep and narrow terrains according to claim 1, wherein in step S7, the earthwork amount calculation based on the collaborative live-action three-dimensional model specifically includes: s7a, acquiring a cooperative live-action three-dimensional model of the same region in the front and rear two stages of construction; S7b, placing the two-stage collaborative live-action three-dimensional model under the same coordinate system, and carrying out fine registration based on stable unchanged ground feature characteristics in the collaborative live-action three-dimensional model, wherein the plane error after registration is less than or equal to 0.03m; S7c, constructing the same regular triangular grids on the surface of the collaborative live-action three-dimensional model, setting the unit side length of the triangular grids to be 5-10 times of the ground sampling distance GSD, and calculating the elevation difference delta H of each unit on the collaborative live-action three-dimensional model in two stages by taking the triangular grids as basic units; And S7d, according to the delta H and the area of all grid cells, summarizing the volume of the delta H >0 part as a filling quantity, and summarizing the volume of the delta H <0 part as a digging quantity.
9. The method for collaborative modeling by using the earthwork oblique photography of the high steep narrow terrain according to claim 1 is characterized in that in the step S7, a construction machine path planning is specifically implemented by guiding a collaborative live-action three-dimensional model into a construction machine control system, planning a transportation path in the collaborative live-action three-dimensional model by the system according to a preset safety rule, wherein the safety rule comprises that the path gradient is less than or equal to 15% and the distance from a dangerous area is more than or equal to 5m, and displaying real-time position and gesture information of the construction machine on the collaborative live-action three-dimensional model in a superimposed mode for construction monitoring and scheduling.
10. The collaborative modeling method for earthmoving oblique photography of high and steep narrow terrain according to claim 9, wherein the step of planning the transportation path further comprises a path smoothing process of smoothing the preliminarily planned path by using spline curves which satisfy a curvature change rate of 0.1/m or less, and deriving the smoothed path coordinates into a construction machine control system compatible format.

Description

Cooperative modeling method for earthwork oblique photography of high-steep narrow terrain Technical Field The invention relates to the technical field of engineering mapping and three-dimensional digital modeling. More particularly, the invention relates to a collaborative modeling method for earthwork oblique photography of high steep narrow terrains. Background In engineering construction of pumped storage power stations, mines, mountain highway slopes, mine dumping sites, valleys and the like, measurement and modeling requirements of high and steep narrow terrains are often met. The terrain has the remarkable characteristics of V-shaped development of the gully, extremely large local gradient (which can exceed 60 degrees), steep mountain and narrow construction area, and brings serious challenges to engineering measurement. Along with the development of mapping technology, unmanned plane laser radar (LiDAR) and oblique photography technology are gradually applied to such scenes, but single technology still has obvious limitations that an unmanned plane is usually in a fixed-height flight mode, is difficult to attach to the fluctuation change of high and steep terrains, is easy to cause image distortion and point cloud deletion of scarp areas, and is insufficient in data acquisition integrity, the laser radar can efficiently acquire high-density terrains and point cloud and penetrate vegetation canopy to directly detect the ground surface, but single-view scanning can cause incomplete side information of the ground surface, scanning blind areas exist, the oblique photography technology can acquire abundant ground surface texture information to construct a three-dimensional model with strong sense of reality, but the generated geometric model has distortion, stretching and other precision problems due to image matching errors in steep and severely shielded areas, and the double requirements of high-precision geometric measurement and complete texture expression cannot be met. Even if two technologies are tried to be combined, the difficulty of collaborative processing of multi-source data still exists that the laser radar and the oblique photographing equipment are mainly independently collected, a unified time synchronization mechanism is lacked, so that obvious space deviation errors exist in data, the subsequent fusion precision is greatly reduced, the point cloud denoising is mainly adopted, non-ground points in a high cliff wall area are easily misjudged to be effective data by adopting a filtering algorithm with fixed parameters, ground points can be mistakenly deleted in a gentle area, filtering adaptability is poor, although the geometric precision of the laser radar point cloud is high, texture information is lacked, the texture of an oblique photographing three-dimensional grid model is rich, the elevation precision is insufficient, the fusion of the laser radar point cloud and the oblique photographing three-dimensional grid model is mainly simple superposition, and collaborative improvement of geometric precision and texture authenticity cannot be realized. How to effectively integrate the high-precision geometric acquisition capability of the laser radar with the rich texture acquisition capability of oblique photography to generate a precise and vivid three-dimensional model, is directly applied to precise calculation of the earthwork and intelligent management and control of construction, and has important significance for improving the efficiency, safety and digitization level of complex terrain engineering construction. Disclosure of Invention The invention provides a collaborative modeling method for earthwork oblique photography of a high and steep narrow terrain, which can be used for earthwork mapping, metering and construction scheduling of the high and steep narrow terrain, and realizes collaborative promotion of modeling geometric accuracy and texture authenticity. To achieve these objects and other advantages and in accordance with the purpose of the invention, there is provided a collaborative modeling method of earthmoving oblique photography of a high steep narrow terrain, comprising the steps of: S1, designing a ground-imitating flight route based on a digital elevation model of a high-steep narrow terrain, controlling an unmanned aerial vehicle carrying a laser radar and at least two oblique angle photographic lenses to fly along the route, and ensuring synchronous execution of laser radar scanning and multi-angle oblique photography through a time synchronization device; S2, arranging image control points in a region, namely arranging a first-stage control point network according to the density of 4-6 photographic baselines at heading intervals and 2 routes at side intervals along a ground-imitating flying route, and adding four-corner control points of an area network to serve as second-stage encryption points aiming at landform abrupt positions or key construction areas; S