Engineering slope three-dimensional surveying method based on photogrammetry

Abstract

This invention relates to the fields of photogrammetry and geographic information services, and discloses a three-dimensional surveying method for engineering slopes based on photogrammetry. The method acquires engineering slope surveying task information and establishes zoning benchmark data, dividing the engineering slope surveying area into slope zoning units and establishing acquisition status records; it acquires first multi-view image data and ground constraint point data, and generates initial point cloud data through image filtering, corresponding point matching, and bundle adjustment; based on the measurability data of the point cloud, it identifies candidate points for supplementary surveys and continuous areas, generates second photogrammetric acquisition parameters, and performs local updates to the zoning, forming a three-dimensional surveying model, slope structure trace results, and a geographic information service package.

Description

Technical Field

[0001] This invention relates to the fields of photogrammetry and geographic information services, and in particular to a three-dimensional surveying method for engineering slopes based on photogrammetry. Background Technology

[0002] Engineering slopes are widely found in mining, highway, railway, water conservancy, and municipal construction projects. Their slopes typically feature multiple slope levels, numerous platforms, significant local obstructions, limited visibility at the slope toe, and uneven distribution of fissures and erosion traces. To obtain survey results on the spatial morphology, slope structure traces, cross-sections, and contour lines of engineering slopes, existing technologies widely employ techniques such as UAV aerial imagery, oblique photogrammetry, multi-view image matching, bundle adjustment, dense point cloud generation, and 3D model construction to assist in the 3D mapping of engineering slopes and the expression of surveying and mapping geographic information service results.

[0003] For example, CN105184863A discloses a method for 3D slope reconstruction based on UAV aerial imagery sequences. This method performs feature region matching and feature point extraction on multi-view UAV aerial imagery sequences, uses bundle adjustment to restore the slope geometry and camera motion parameters, obtains a sparse slope 3D point cloud model, then uses a multi-view stereo vision algorithm to diffuse it into a dense slope 3D point cloud model, and further performs surface mesh reconstruction and texture mapping to establish a 3D slope model. This existing technology can achieve the reconstruction of 3D slope models and is a photogrammetric slope 3D modeling scheme that is quite similar to this application.

[0004] However, in the actual surveying of engineering slopes, insufficient image visibility, unstable ray intersection angles, local point cloud discontinuities, or insufficient texture support are common problems in areas such as the slope toe, lower edge of the platform, concave slope surfaces, vegetation-covered areas, and weakly textured slope surfaces. The existing technologies mentioned above mainly focus on the reconstruction process of generating sparse point clouds, dense point clouds, and 3D models from multi-view images. They have not yet established a zoning judgment mechanism for the measurability differences of different slope surface units in engineering slopes, and it is difficult to determine the areas requiring supplementary surveying based on initial camera pose data, point cloud discontinuity status, and texture support status.

[0005] Therefore, in the existing process of 3D surveying and mapping geographic information services for engineering slopes, for point cloud missing areas caused by occlusion, weak texture or insufficient viewpoint, the supplementary survey area and direction often need to be judged by human experience. It is difficult to establish a stable correspondence between the missing state in the initial 3D reconstruction results and the subsequent supplementary survey parameters, which in turn affects the zonal updating of engineering slope 3D survey results and the organization of mapping geographic information service results.

[0006] Therefore, the technical problem that the existing technology urgently needs to solve is: how to identify the measurability of point cloud missing areas caused by occlusion, weak texture and insufficient viewpoint during the three-dimensional survey of engineering slope photogrammetry, and generate directional supplementary measurement acquisition parameters based on the identification results, so as to support the zonal updating of engineering slope three-dimensional survey results and the output of surveying and mapping geographic information service results. Summary of the Invention

[0007] To overcome the aforementioned technical deficiencies, the present invention aims to provide a three-dimensional surveying method for engineering slopes based on photogrammetry. By establishing slope zoning units and acquisition status records, measurability data of the point cloud is generated based on initial camera pose data, initial dense point cloud data of the engineering slope, and acquisition status records. Candidate points for supplementary measurement and their continuous regions are identified, and second photogrammetric acquisition parameters are generated accordingly to perform localized updates of the initial dense point cloud data of the engineering slope. This solves the problem of difficulty in directional supplementary measurement and zonal updates of missing point cloud areas caused by occlusion, weak texture, and insufficient viewing angle in three-dimensional photogrammetric surveying of engineering slopes.

[0008] This invention discloses a three-dimensional surveying method for engineering slopes based on photogrammetry, comprising the following steps: S1. Obtain engineering slope survey task information and establish zoning benchmark data for the engineering slope survey area. Engineering slope survey task information includes surveying and mapping geographic information service task identifiers. Zoning benchmark data includes slope top line, slope toe line, platform line, obstruction boundary, existing ground constraint points, and existing slope image data. S2. Based on the slope top line, slope toe line, platform line and obstruction boundary, the engineering slope survey area is divided into several slope partition units, and a data acquisition status record is established for each slope partition unit. The data acquisition status record includes the slope partition unit number, target ground resolution, candidate direction of structural trace, obstruction risk level and supplementary measurement trigger conditions. S3. Generate the first photogrammetric acquisition parameters according to the acquisition status record, and acquire the first multi-view image data and ground constraint point data of the engineering slope survey area according to the first photogrammetric acquisition parameters. The first multi-view image data includes orthophoto image, oblique image and slope toe lateral image. S4. Perform image filtering, corresponding point matching and bundle adjustment on the first multi-view image data to generate initial camera pose data, initial sparse point cloud data and initial engineering slope dense point cloud data. S5. Based on the initial camera pose data, the initial dense point cloud data of the engineering slope, and the acquisition status record, generate the point cloud measurability data for each slope partition unit. The point cloud measurability data includes the number of times the image is visible, the range of ray intersection angles, the local point cloud discontinuity boundaries, and the texture support status. S6. Based on the measurability data of the point cloud, the initial dense point cloud data of the engineering slope is classified into initial surface points, initial slope structure trace points and supplementary measurement candidate points. When the supplementary measurement candidate points in the same slope partition unit form a continuous area and meet the supplementary measurement triggering conditions, the second photogrammetric acquisition parameters are generated according to the location, boundary, slope partition unit number to which the continuous area belongs, and the initial camera pose data corresponding to the continuous area. S7. Acquire second multi-view image data according to the second photogrammetric acquisition parameters. The second multi-view image data includes at least one of supplementary oblique view image and supplementary slope toe lateral view image. Incorporate the second multi-view image data into the corresponding slope partition unit. Perform partitioned local updates on the initial engineering slope dense point cloud data to obtain updated engineering slope dense point cloud data. Reclassify the updated engineering slope dense point cloud data to obtain updated surface points and updated slope structure trace points. S8. Construct a three-dimensional survey model of the engineering slope based on the updated surface points, generate slope structure trace results based on the updated slope structure trace points, and generate a surveying and mapping geographic information service result package based on the three-dimensional survey model of the engineering slope and the slope structure trace results. The surveying and mapping geographic information service result package includes the three-dimensional survey model of the engineering slope, orthophoto, point cloud file, cross-section results, contour line results, slope structure trace results, supplementary survey record file and metadata file.

[0009] Preferably, in step S2, when dividing the slope into partition units, the longitudinal zone of the slope is first determined according to the top line, bottom line and platform line, and the transverse segment of the slope is determined according to the shading boundary and existing ground constraint points. The longitudinal zone and transverse segment of the slope are then superimposed to form the slope partition unit.

[0010] Preferably, in step S2, the candidate direction of the structural trace is determined based on the linear texture direction of the existing slope image data within the slope partition unit, the direction of the platform line adjacent to the slope partition unit, and the direction of the slope toe line adjacent to the slope partition unit. When the angle between the linear texture direction of the existing slope image data within the slope partition unit and the direction of the platform line adjacent to the slope partition unit or the direction of the slope toe line adjacent to the slope partition unit is within a preset angle range, the linear texture direction is written into the acquisition status record.

[0011] Preferably, in step S3, when generating the first photogrammetric acquisition parameters, orthophoto acquisition sub-parameters, tilt acquisition sub-parameters, and slope toe lateral acquisition sub-parameters are set for each slope partition unit. The tilt acquisition sub-parameter sets the camera orientation based on the candidate direction of the structural trace, while the slope foot lateral acquisition sub-parameter sets the acquisition distance and camera elevation angle based on the occlusion risk level.

[0012] Preferably, in step S3, the ground constraint point data includes partition boundary constraint points, slope transition constraint points, and independent check points; The boundary constraint points of the zoning are set at the junction of adjacent slope zoning units, and the slope transition constraint points are set at the top line, platform line and toe line of the slope. Independent checkpoints do not participate in the bundle adjustment.

[0013] Preferably, in step S3, when acquiring the first multi-view image data, the image view type, slope partition unit number, camera orientation, acquisition height, acquisition distance and acquisition time are written into the image metadata of the corresponding image; Image perspective types include orthographic perspective, oblique perspective, and slope-foot lateral perspective.

[0014] Preferably, in step S4, when performing corresponding point matching, a partitioned image association table is established based on image metadata, and within the same slope partition unit, corresponding points between images of different types in the orthophoto view type, oblique view type, and slope foot lateral view type are matched first. When the number of points with the same name within the same slope partition unit is less than the preset number, the slope partition unit is added to the list of partitions to be evaluated.

[0015] Preferably, in step S5, when generating point cloud measurability data, the correspondence between image rays and point cloud points in the initial engineering slope dense point cloud data is established based on the initial camera pose data, and the number of times the image is visible and the range of ray intersection angles corresponding to each point cloud point are counted. When the number of visible images is lower than the preset number of visible images or the ray intersection angle range does not meet the preset intersection angle range, the corresponding point cloud points are marked as low measurability points.

[0016] Preferably, in step S6, when classifying the initial dense point cloud data of the engineering slope into initial surface points, initial slope structure trace points, and supplementary measurement candidate points, the following classification rules are adopted: Point cloud points that are spatially continuous and whose local flatness meets the preset flatness conditions are identified as initial surface points. Point cloud points whose texture support state satisfies the linear texture condition, have normal variation boundaries, and cross-image visibility are determined as the initial slope structure trace points. Points that are low measurability points and are located near the discontinuity boundary of the local point cloud are identified as candidate points for supplementary measurement.

[0017] Preferably, in step S6, when generating the second photogrammetric acquisition parameters, the supplementary measurement center is determined based on the location of the continuous area, the supplementary measurement coverage is determined based on the boundary of the continuous area, the existing acquisition blind direction is determined based on the initial camera pose data corresponding to the continuous area, and the acquisition direction that avoids the existing acquisition blind direction is determined as the supplementary measurement acquisition direction.

[0018] Preferably, in step S7, when performing local updates of the initial engineering slope dense point cloud data, the slope partition units that do not meet the supplementary measurement triggering conditions are locked, and local dense matching is performed on the slope partition units that meet the supplementary measurement triggering conditions and their adjacent slope partition units. After local dense matching is completed, the point cloud boundaries between adjacent slope partition units are stitched together based on the partition boundary constraint points.

[0019] Preferably, in step S8, when generating the slope structure trace results, the updated slope structure trace points are aggregated to obtain crack traces, slope scour traces, and step boundary traces. The fissure traces, slope scour traces, and step boundary traces were written into different layers of the slope structure trace results.

[0020] Preferably, in step S8, when generating the cross-section results, the key cross-section locations are determined based on the spatial density of the updated slope structure trace points, and transverse cross-section lines are generated at the key cross-section locations. A longitudinal section line is generated at the intersection of two adjacent slope partition units. Both the transverse and longitudinal section lines intersect with the three-dimensional survey model of the engineering slope to generate section results.

[0021] Preferably, in step S8, when generating the metadata file, the mapping geographic information service task identifier, slope zoning unit number, first photogrammetry acquisition parameters, second photogrammetry acquisition parameters, supplementary survey triggering conditions, continuous area boundaries, zoning local update records, and result file index are written into the metadata file.

[0022] Preferably, after step S8, the method further includes performing a partition consistency verification on the surveying and mapping geographic information service result package; The zoning consistency verification includes: verifying the spatial position deviation of the corresponding slope zoning unit based on independent checkpoints, verifying the boundary splicing deviation of adjacent slope zoning units based on zoning boundary constraint points, and writing the spatial position deviation and boundary splicing deviation into the supplementary measurement record file.

[0023] Compared with existing technologies, the above technical solution has the following advantages: 1. This invention establishes slope zoning units and acquisition status records, and associates the slope zoning unit number, structural trace candidate direction, occlusion risk level, and supplementary measurement trigger conditions in the acquisition status records. This enables the three-dimensional surveying process of engineering slopes to be zoned according to the visibility conditions, occlusion conditions, and structural characteristics of different slope areas, which is beneficial to improving the pertinence of engineering slope photogrammetry surveys.

[0024] 2. This invention generates point cloud measurability data based on initial camera pose data, initial dense point cloud data of engineering slopes, and acquisition status records. It identifies low measurability areas by the number of times images are visible, the range of ray intersection angles, the local point cloud discontinuity boundaries, and the texture support status. This allows for the quantitative judgment of point cloud missing areas caused by occlusion, weak texture, or insufficient viewing angle, reducing the reliance on manual experience to determine supplementary measurement areas.

[0025] 3. This invention classifies the initial dense point cloud data of the engineering slope into initial surface points, initial slope structure trace points, and supplementary measurement candidate points. When the supplementary measurement candidate points form a continuous area and meet the supplementary measurement triggering conditions, a second photogrammetric acquisition parameter is generated, enabling the supplementary measurement acquisition to be generated directionally around the actual missing area, the boundary of the continuous area, and the corresponding initial camera pose data, thereby improving the targeting of the supplementary measurement acquisition.

[0026] 4. This invention generates second photogrammetric acquisition parameters based on the location, boundary, slope partition unit number, and initial camera pose data corresponding to the continuous area, and makes the supplementary acquisition direction avoid the existing blind direction. It can supplement the missing or insufficient observation angles in the initial acquisition process, which is beneficial to improving the point cloud integrity in the slope foot area, the lower edge of the platform area, the concave slope area, and the area near the shading boundary.

[0027] 5. After acquiring the second multi-view image data, the present invention performs local updates on the initial dense point cloud data of the engineering slope in partitions, instead of repeatedly performing global dense matching on the entire engineering slope survey area. This can reduce the amount of redundant processing and help maintain the spatial continuity between the supplemented and non-supplemented survey areas.

[0028] 6. By setting partition boundary constraint points, slope transition constraint points, and independent checkpoints, this invention enables ground constraint point data to serve the splicing of adjacent slope partition units, transition constraints of slope top line, platform line, and slope toe line, as well as independent accuracy verification, which helps to improve the spatial consistency of the three-dimensional survey model of engineering slopes and the reliability of result verification.

[0029] 7. This invention generates slope structure trace results based on updated slope structure trace points, and writes crack traces, slope scour traces and step boundary traces into different layers respectively, so that the structural mapping results of engineering slopes can form a corresponding relationship with the three-dimensional survey model, cross-section results and contour line results, which is beneficial to subsequent engineering analysis and result retrieval.

[0030] 8. This invention determines the location of key sections based on the spatial density of updated slope structure trace points, and generates longitudinal section lines by combining the intersection of two adjacent slope partition units. This allows the section results to more centrally reflect the slope structure change area and partition boundary area, which is beneficial to improving the engineering applicability of the engineering slope section results.

[0031] 9. This invention generates contour line results based on updated surface points or digital surface models constructed from updated surface points, and establishes a correlation between the contour line results and slope zoning unit numbers, so that the contour line results can be consistent with slope zoning units, cross-sectional results and slope structure trace results, which is conducive to forming a unified results organization for surveying and mapping geographic information services.

[0032] 10. This invention generates a mapping and geographic information service result package including a 3D survey model of an engineering slope, orthophotos, point cloud files, cross-section results, contour line results, slope structure trace results, supplementary survey record files, and metadata files. The metadata files record the mapping and geographic information service task identifier, slope zoning unit number, first photogrammetry acquisition parameters, second photogrammetry acquisition parameters, supplementary survey triggering conditions, boundaries of continuous areas, zoning local update records, and result file indexes. This is beneficial for improving the organization, traceability, and efficiency of subsequent service calls for the 3D survey results of engineering slopes. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating a three-dimensional surveying method for engineering slopes based on photogrammetry. Figure 2 This is a schematic diagram of the slope zoning unit and multi-view image acquisition in this invention; Figure 3 This is a schematic diagram of the point cloud measurability evaluation and directional supplementary measurement process in this invention; Figure 4 This is a schematic diagram of the organizational structure of the surveying and mapping geographic information service deliverables package in this invention; Figure 5 This is a graph showing the change in point cloud integrity rate with the number of retesting rounds in this invention. Figure 6 This is a schematic diagram of the error comparison of independent checkpoints in this invention; Figure 7 This is a schematic diagram of the geometric calculation of point cloud measurability in this invention. Detailed Implementation

[0034] The advantages of the present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments.

[0035] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0036] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0037] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0038] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0039] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0040] In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the convenience of the description of the invention and have no specific meaning in themselves. Therefore, "module" and "part" can be used interchangeably.

[0041] The invention will be further described below with reference to the accompanying drawings. This invention relates to the fields of photogrammetry and geographic information services, and is particularly applicable to multi-view image acquisition, point cloud measurability evaluation, directional supplementary surveying, zonal local updates, and output of geographic information service packages during the three-dimensional surveying of engineering slopes. It belongs to the application direction of photogrammetry and mapping.

[0042] like Figure 1As shown, this embodiment provides a three-dimensional surveying method for engineering slopes based on photogrammetry. The method takes engineering slope surveying task information as input and outputs a three-dimensional surveying model of the engineering slope, orthophotos, point cloud files, cross-section results, contour line results, slope structure trace results, supplementary survey record files, and metadata files. Instead of simply generating a three-dimensional model from multi-view images, this method generates point cloud measurability data after initial photogrammetric reconstruction based on initial camera pose data, initial dense point cloud data of the engineering slope, and acquisition status records. It further identifies candidate points for supplementary surveys and their continuous regions, and generates second photogrammetric acquisition parameters based on the location, boundaries, slope zoning unit number, and initial camera pose data corresponding to the continuous regions. This enables directional supplementary surveys and local updates of zoning areas with occlusion, weak texture, and insufficient viewing angles.

[0043] In this embodiment, the engineering slope can be a mining slope, highway slope, railway slope, water conservancy project slope, municipal foundation pit slope, or other engineering slopes that require 3D mapping. Engineering slopes typically have characteristics such as multiple slope levels, multiple platforms, concave slope toe, local vegetation obstruction, uneven slope texture, and irregular distribution of crack traces. For this type of engineering slope, if only one-time multi-view image acquisition and conventional 3D reconstruction methods are used, local point cloud discontinuities are easily formed in the slope toe area, the lower edge of the platform area, near the obstruction boundary, and in the weak texture slope area. Therefore, this embodiment uses a process of "slope zoning unit - acquisition status record - point cloud measurability data - supplementary measurement candidate points - second photogrammetric acquisition parameters - zoning local update - surveying and mapping geographic information service result package" to enable photogrammetric 3D surveying to form a measurability judgment and supplementary measurement feedback process.

[0044] In this embodiment, the engineering slope survey task information is first acquired, and zoning benchmark data for the engineering slope survey area is established. The engineering slope survey task information includes a surveying and mapping geographic information service task identifier, which is used to identify this survey task in subsequent surveying and mapping geographic information service deliverable packages, ensuring that the engineering slope 3D survey model, point cloud files, cross-section results, slope structure trace results, supplementary survey record files, and metadata files can all be mapped to the same surveying and mapping geographic information service task. Zoning benchmark data includes the slope crest line, slope toe line, platform line, occlusion boundary, existing ground constraint points, and existing slope imagery data. The slope crest line represents the spatial location of the upper edge of the engineering slope, the slope toe line represents the spatial location of the lower edge of the engineering slope, the platform line represents the platform boundary between multiple slope levels, and the occlusion boundary represents the boundary of locally invisible areas caused by vegetation, structures, equipment, steep slopes, or concave slopes. Existing ground constraint points can be derived from previous measurement results, preliminary field measurement data, or existing control point data. Existing slope image data can be derived from historical aerial images, pre-acquired images, or on-site reconnaissance images, and can be used to determine the direction of linear texture and candidate direction of structural traces on the slope before formal acquisition.

[0045] like Figure 2 As shown, after establishing the zoning benchmark data, the engineering slope survey area is divided into several slope zoning units based on the slope crest line, slope toe line, platform line, and shading boundary. When dividing the slope zoning units, the longitudinal zones of the slope are first determined based on the slope crest line, slope toe line, and platform line. Then, the transverse segments of the slope are determined based on the shading boundary and existing ground constraint points. The longitudinal zones and transverse segments are then overlaid to form the slope zoning units. In this way, each slope zoning unit has a clear spatial boundary and surveying object, facilitating subsequent establishment of data acquisition status records, statistical analysis of point cloud measurability data, identification of candidate points for supplementary surveys, and execution of local zoning updates. Unlike ordinary rectangular grid zoning, the slope zoning units in this embodiment are formed by combining the slope crest line, slope toe line, platform line, shading boundary, and existing ground constraint points, which is beneficial for adapting to changes in slope grade, platform, shading, and slope toe of the engineering slope.

[0046] Subsequently, an acquisition status record was established for each slope partition unit. The acquisition status record includes the slope partition unit number, target ground resolution, candidate structural trace direction, occlusion risk level, and re-measurement trigger condition. The slope partition unit number is used to uniquely identify the corresponding slope partition unit; the target ground resolution is used to limit the image acquisition detail of the slope partition unit; the candidate structural trace direction is used to characterize the possible direction of the extension of cracks, slope erosion traces, or step boundaries within the slope partition unit; the occlusion risk level is used to indicate the degree of occlusion or point cloud omissions that may occur during photogrammetric acquisition of the slope partition unit; and the re-measurement trigger condition is used to determine whether a second multi-view image data acquisition and local partition update are needed for the slope partition unit.

[0047] The candidate direction of the structural trace is determined based on the linear texture direction of the existing slope image data within the slope partition unit, the direction of the platform line adjacent to the slope partition unit, and the direction of the slope toe line adjacent to the slope partition unit. When the angle between the linear texture direction of the existing slope image data within the slope partition unit and the direction of the platform line or the slope toe line adjacent to the slope partition unit is within a preset angle range, the linear texture direction is written into the acquisition status record. The preset angle range can be set according to the engineering slope type and surveying accuracy requirements, for example, from 15 degrees to 75 degrees. Through this processing, the possible directions of the slope structural trace can be obtained in advance before the formal acquisition, providing a basis for setting the camera orientation in the subsequent tilt acquisition sub-parameters.

[0048] The first photogrammetric acquisition parameters are generated based on the acquisition status record, and first multi-view image data and ground constraint point data of the engineering slope survey area are acquired according to the first photogrammetric acquisition parameters. The first multi-view image data includes orthophoto images, oblique perspective images, and slope toe lateral perspective images. Orthophoto images are used to obtain the overall spatial relationship of the slope and orthophoto results; oblique perspective images are used to enhance the intersection conditions of the slope's three-dimensional reconstruction; slope toe lateral perspective images are used to supplement the visibility of the slope toe area, the lower edge of the platform area, and the concave slope area. When generating the first photogrammetric acquisition parameters, orthophoto acquisition sub-parameters, oblique acquisition sub-parameters, and slope toe lateral acquisition sub-parameters are set for each slope partition unit. The oblique acquisition sub-parameters set the camera orientation according to the candidate direction of the structural trace, so that the camera acquisition direction can cover the direction in which the structural trace may extend. The slope toe lateral acquisition sub-parameters set the acquisition distance and camera elevation angle according to the occlusion risk level, so that the slope toe lateral perspective images can cover the concave area at the slope toe that is difficult to cover by conventional orthophoto and oblique perspective images.

[0049] Ground constraint point data includes zoning boundary constraint points, slope transition constraint points, and independent checkpoints. Zoning boundary constraint points are located at the boundaries of adjacent slope zoning units and are used to control the point cloud boundary stitching between adjacent slope zoning units after local zoning updates. Slope transition constraint points are located at the top, plateau, and toe lines of the slope and are used to maintain spatial constraints on the upper and lower boundaries of the slope and the transition positions of the plateau. Independent checkpoints do not participate in bundle adjustment and are used for subsequent zoning consistency verification of the surveying and mapping geographic information service deliverables. By classifying ground constraint point data into these three categories, different points can serve zoning stitching, slope transition constraints, and independent accuracy verification respectively.

[0050] When acquiring the first multi-view image data, the image viewpoint type, slope partition unit number, camera orientation, acquisition height, acquisition distance, and acquisition time are written into the corresponding image metadata. Image viewpoint types include orthophoto viewpoint, oblique viewpoint, and slope toe lateral viewpoint. After establishing a link between the image metadata and the slope partition unit number, subsequent matching of corresponding points can prioritize matching corresponding points between images of different types (orthophoto viewpoint, oblique viewpoint, and slope toe lateral viewpoint) within the same slope partition unit, thereby forming a stable cross-viewpoint constraint relationship between images from different viewpoints.

[0051] After acquiring the first multi-view image data, image filtering, corresponding point matching, and bundle adjustment are performed on the first multi-view image data to generate initial camera pose data, initial sparse point cloud data, and initial dense point cloud data of the engineering slope. Image filtering can be performed based on image sharpness, exposure status, overlap relationship, visibility of ground constraint points, and integrity of image metadata. In this embodiment, the first multi-view image data retained after image filtering and participating in corresponding point matching and bundle adjustment is called valid image. When performing corresponding point matching, a partitioned image association table is established based on image metadata, and within the same slope partition unit, corresponding points between images of different types (orthogonal view type, oblique view type, and slope toe lateral view type) are matched first. When the number of corresponding points within the same slope partition unit is less than a preset number, the slope partition unit is added to the partition list to be evaluated. The slope partition units in the partition list to be evaluated are the focus of inspection in the subsequent point cloud measurability evaluation to determine whether there are insufficient image visibility counts, insufficient ray intersection angle ranges, or local point cloud discontinuities.

[0052] Bundle adjustment uses the corresponding point matching results, the partition boundary constraint points and slope transition constraint points in the ground constraint point data, and the first photogrammetric acquisition parameters to obtain initial camera pose data. The initial camera pose data includes the camera position and camera orientation in the photogrammetric coordinate system for each valid image. Initial sparse point cloud data is generated based on the initial camera pose data and the corresponding point matching results, and then initial dense point cloud data for the engineering slope is generated through multi-view dense matching. The initial dense point cloud data for the engineering slope is the foundational data for subsequent point cloud measurability evaluation, supplementary measurement candidate point identification, and 3D survey model construction.

[0053] like Figure 7As shown, after obtaining the initial dense point cloud data of the engineering slope, the measurability data of the point cloud for each slope partition unit is generated based on the initial camera pose data, the initial dense point cloud data of the engineering slope, and the acquisition status records. The measurability data of the point cloud includes the number of times the image is visible, the range of ray intersection angles, the local point cloud discontinuity boundaries, and the texture support status. When generating the point cloud measurability data, the correspondence between image rays and point cloud points in the initial dense point cloud data of the engineering slope is established based on the initial camera pose data, and the number of times the image is visible and the range of ray intersection angles corresponding to each point cloud point are counted. The number of times the image is visible indicates how many effective images can observe a certain point cloud point; the range of ray intersection angles indicates the intersection angle conditions formed between different image observation rays participating in the reconstruction of the point cloud point; the local point cloud discontinuity boundaries indicate the boundary positions where local missing or density abrupt changes occur in the continuous surface of the point cloud; and the texture support status indicates whether the image area near the point cloud point has sufficient texture features to support stable matching.

[0054] Texture support status can be determined based on the gray-level gradient, edge continuity, texture repetition, and feature point distribution density within the image neighborhood corresponding to the point cloud points. Gray-level gradient reflects the degree of brightness variation within the image neighborhood; edge continuity reflects whether linear textures have continuous extension characteristics; texture repetition reflects the presence of repeated textures in the image neighborhood—high repetition reduces the stability of matching corresponding points; and feature point distribution density reflects the degree of local feature support required for matching corresponding points. Local point cloud discontinuities can be determined by statistically analyzing abrupt changes in distance between adjacent point cloud points, abrupt changes in local point density, and abrupt changes in the side length of the triangulation mesh. When both insufficient texture support and local point cloud discontinuities occur simultaneously in the same region, this region is more likely to be identified as a candidate region for supplementary measurements.

[0055] In a specific calculation example, let point cloud point P be a point cloud point that is observed by 4 valid images. Then, the number of times point cloud point P is visible in the images is calculated. for: Among them, the number of times the image is visible This represents the number of valid images that can be observed at point P in the point cloud. If the valid intersection angles formed between the observed rays corresponding to the four valid images are 18 degrees, 31 degrees, 44 degrees, and 57 degrees, respectively, then the lower limit of the ray intersection angle for point P in the point cloud is... Upper limit of the intersection angle of rays They are respectively: Among them, the lower limit of the ray intersection angle The minimum effective intersection angle among those involved in the reconstruction of point cloud point P is represented by the upper limit of the ray intersection angle. This represents the maximum effective intersection angle among those involved in the reconstruction of point P in the point cloud. If the preset number of visible points is 3, and the preset intersection angle range is 15 degrees to 70 degrees, then due to the number of image visible points... Equal to 4th order, lower limit of ray intersection angle Equal to 18 degrees and the upper limit of the ray intersection angle The point P in the point cloud is equal to 57 degrees and satisfies the measurability condition, so it is not marked as a low measurability point.

[0056] In another specific calculation example, suppose point Q in the point cloud is observed by only 2 valid images, and calculate the number of times point Q is visible in the images. for: If the effective intersection angle corresponding to point Q in the point cloud is 9 degrees, then the lower limit of the ray intersection angle of point Q in the point cloud is... Upper limit of the intersection angle of rays Both are 9 degrees. (Due to the number of times the image is visible...) The number of visible occurrences is less than 3 times, and the lower limit of the ray intersection angle is also lower. Point Q is marked as a low measurability point because it is 15 degrees below the lower limit of the preset intersection angle range. If point Q is also located near the local point cloud discontinuity boundary and the texture support is insufficient, then point Q is further identified as a candidate point for supplementary measurement.

[0057] Based on the measurability data of point clouds, the initial dense point cloud data of the engineering slope is classified into initial surface points, initial slope structure trace points, and supplementary measurement candidate points. During classification, point cloud points that are spatially continuous and whose local flatness meets preset flatness conditions are identified as initial surface points. Point cloud points whose texture support state meets linear texture conditions, have normal variation boundaries, and cross-image visibility are identified as initial slope structure trace points. Point cloud points that are low measurability points and located near local point cloud discontinuity boundaries are identified as supplementary measurement candidate points. Initial surface points are used to construct the main ground surface of the 3D survey model of the engineering slope; initial slope structure trace points are used to identify crack traces, slope erosion traces, and step boundary traces; supplementary measurement candidate points are used to determine whether directional supplementary measurement is needed.

[0058] The normal variation boundary can be determined based on the angle between the normals of adjacent local fitted planes within the neighborhood of a point cloud point. Specifically, adjacent local planes can be fitted within the neighborhood of a point cloud point, and the angle between the normals of these adjacent local planes can be calculated. When this angle is greater than a preset normal variation threshold, the corresponding region is determined as the normal variation boundary. Cross-image visibility can be determined based on whether the same point cloud point has matching points in images corresponding to different image viewpoint types. For example, if a point cloud point has corresponding matching points in at least two types of images—orthophoto view, oblique view, or slope foot lateral view—it can be considered to have cross-image visibility. By jointly judging the normal variation boundary and cross-image visibility, the initial slope structure trace points can be more concentrated in areas with structural abrupt changes and cross-view image support.

[0059] To make the supplementary measurement trigger judgment clearer, this implementation method can determine the trigger based on the continuous area formed by the candidate supplementary measurement points within the same slope zoning unit and the proportion of low measurability points. Let the continuous area formed by the candidate supplementary measurement points within a certain slope zoning unit be... The threshold for the triggered area in the supplementary test is The number of low measurability points in a continuous region is The total number of point cloud points in the continuous region is The proportion of low measurability points is The proportion of low measurability points Calculate according to the following standard mathematical formula: Among them, the number of low measurability points in a continuous region This represents the number of point cloud points marked as low measurability points in a continuous region; the total number of point cloud points in a continuous region. This represents the total number of point cloud points in a continuous region; the proportion of points with low measurability. This represents the proportion of low measurability points to the total number of points in the point cloud within a continuous region.

[0060] In a specific example, the area of ​​the continuous region formed by the candidate points for supplementary measurement within a certain slope zoning unit. The area is 32 square meters; the trigger area threshold needs to be retested. The area is 20 square meters; the number of low measurability points within a continuous area. The total number of point cloud points in a continuous area is 1200. If there are 2000, then the proportion of low measurability points is... for: If the trigger condition for the supplementary test is set to the area of ​​a continuous region. Not less than 20 square meters and low measurability point ratio If the percentage is not less than 50%, then the continuous area meets the trigger condition for supplementary measurement, and a second photogrammetric acquisition parameter needs to be generated. This judgment method can avoid frequent supplementary measurements for scattered noise points or small error areas, and concentrate the supplementary measurement process on low measurability areas with practical mapping significance.

[0061] When candidate points for supplementary measurements within the same slope zoning unit form a continuous area and meet the supplementary measurement triggering conditions, second photogrammetric acquisition parameters are generated based on the location, boundary, slope zoning unit number, and initial camera pose data corresponding to the continuous area. The continuous area can be determined by the spatial connectivity, boundary closure, and area range of the candidate points for supplementary measurements. When generating the second photogrammetric acquisition parameters, the supplementary measurement center is determined based on the location of the continuous area, the supplementary measurement coverage is determined based on the boundary of the continuous area, and the existing blind direction is determined based on the initial camera pose data corresponding to the continuous area. This blind direction refers to the observation direction that results in a low visibility count or does not meet the ray intersection angle requirement for the continuous area during the initial acquisition process, and the acquisition direction that avoids the existing blind direction is determined as the supplementary measurement acquisition direction.

[0062] Existing blind directions can be statistically determined based on the initial camera pose data corresponding to low measurability points within a continuous area. When the number of visible images in a certain observation direction within a continuous area is lower than the preset number of visible images, or the corresponding ray intersection angle range does not meet the preset intersection angle range, that observation direction is determined as an existing blind direction. The supplementary acquisition direction in the second photogrammetric acquisition parameters avoids existing blind directions to supplement the missing or insufficient observation angles during the initial acquisition process. Thus, the second multi-view image data can supplement the insufficient observation directions in the initial acquisition, rather than simply repeating the original acquisition direction.

[0063] In an example of calculating second photogrammetric acquisition parameters, the minimum circumscribed rectangle of the continuous area is assumed to be 24 meters long and 12 meters wide. To ensure that the supplementary survey coverage extends beyond the transition area outside the continuous area boundary, a 3-meter extension distance is set outside the continuous area boundary. Therefore, the supplementary survey coverage length... and supplementary coverage width They are respectively: Among them, the supplementary measurement coverage length The measurement indicates the coverage length along the length of the continuous area, in meters; the measurement also indicates the coverage width. This represents the coverage width along the width of the continuous area during the supplementary survey, in meters. If the ground resolution of the target area is 2 centimeters, the ground coverage width of a single image from the supplementary camera is 10 meters, and the lateral overlap rate is set to 60%, then the effective lateral spacing between adjacent supplementary images is... for: Here, 10 meters represents the ground coverage width of a single image from the supplementary measurement camera, and the effective lateral spacing. This indicates the effective coverage spacing between adjacent supplementary survey images in the lateral direction, in meters; a value of 0.60 indicates a lateral overlap rate of 60%. To cover the supplementary survey coverage width... In an area equal to 18 meters, the number of lateral sampling strips was measured again. The following method can be used to estimate: Among them, the number of lateral sampling strips was supplemented. Indicates the number of data strips required to complete the supplementary measurement coverage width, symbol [symbol missing]. This indicates rounding up. Therefore, the supplementary coverage length in the second photogrammetric acquisition parameters is 30 meters, the supplementary coverage width is 18 meters, the effective lateral spacing is 4 meters, and the number of supplementary lateral acquisition strips is 6. If the initial camera pose data indicates that the continuous area mainly lacks the lateral observation direction at the toe of the slope, the second photogrammetric acquisition parameters will also set the supplementary acquisition direction from the lateral direction at the toe of the slope towards the slope surface, and configure supplementary lateral view images at the toe of the slope.

[0064] It should be noted that the preset angle range, preset visibility count, preset intersection angle range, supplementary measurement trigger area threshold, low measurability point proportion threshold, supplementary measurement outward extension distance, target ground resolution, image overlap rate, and allowable error range involved in this embodiment are all exemplary parameters used to illustrate the calculation and implementation process. In actual engineering slope 3D surveying, the above parameters can be set according to the scale of the engineering slope, slope complexity, mapping scale, camera parameters, surveying accuracy requirements, and surveying and mapping geographic information service results requirements.

[0065] Second multi-view image data is acquired according to the second photogrammetric acquisition parameters. The second multi-view image data includes at least one of supplementary oblique view images and supplementary slope toe lateral view images. For slope toe areas or platform lower edge areas, supplementary slope toe lateral view images can be acquired first; for slope areas with obvious cracks or scour traces but insufficient initial texture support, supplementary oblique view images can be acquired first. The second multi-view image data is incorporated into the corresponding slope partition units, and the initial engineering slope dense point cloud data is locally updated in partitions to obtain updated engineering slope dense point cloud data. During the local partition update, slope partition units that do not meet the supplementary measurement trigger conditions are locked, and local dense matching is performed on slope partition units that meet the supplementary measurement trigger conditions and their adjacent slope partition units. After the local dense matching is completed, the point cloud boundaries between adjacent slope partition units are stitched together based on the partition boundary constraint points. This method can reduce the need to repeatedly perform global dense matching on the entire engineering slope survey area and can maintain the spatial continuity between the supplementary and non-supplementary areas.

[0066] After completing the local update of the partitioned area, the dense point cloud data of the updated engineering slope is reclassified to obtain updated surface points and updated slope structure trace points. The updated surface points are used to construct a 3D survey model of the engineering slope. The updated slope structure trace points are used to generate slope structure trace results. When constructing the 3D survey model of the engineering slope based on the updated surface points, a digital surface model of the engineering slope can be generated first, then a triangular mesh surface can be generated based on the spatial relationships of the point cloud, and the effective image texture can be mapped onto the triangular mesh surface to obtain a 3D survey model of the engineering slope that can be used to express the results of surveying and mapping geographic information services. When generating slope structure trace results based on the updated slope structure trace points, line segment aggregation is performed on the updated slope structure trace points to obtain fissure traces, slope erosion traces, and step boundary traces. These fissure traces, slope erosion traces, and step boundary traces are then written into different layers of the slope structure trace results.

[0067] When generating cross-sectional results, the location of key cross-sections is determined based on the spatial density of updated slope structure trace points, and transverse cross-sectional lines are generated at these key cross-sectional locations. The spatial density of updated slope structure trace points can be determined by the number of updated slope structure trace points per unit area. When the number of updated slope structure trace points per unit area exceeds a preset density threshold, the corresponding area is designated as a key cross-sectional location. For slope areas with densely distributed updated slope structure trace points, it can be assumed that the area contains a lot of structural information such as cracks, scour, or step boundaries, requiring the generation of key cross-sectional locations to represent the slope geometry changes in that area. Longitudinal cross-sectional lines are generated based on the intersection of two adjacent slope partition units. Both transverse and longitudinal cross-sectional lines intersect with the 3D survey model of the engineering slope to generate cross-sectional results. Compared with generating cross-sections at fixed intervals, this method allows the cross-sectional results to be more concentrated in the structural change areas and partition boundary areas of the engineering slope.

[0068] When generating contour line results, contour lines can be extracted from updated surface points in the 3D survey model of the engineering slope or from a digital surface model constructed using updated surface points, according to preset elevation intervals. The contour line results are then linked to slope zoning unit numbers and written into the surveying and mapping geographic information service deliverables package. The preset elevation intervals can be set according to the engineering slope mapping scale, slope elevation difference, and surveying and mapping geographic information service deliverables requirements. By linking contour line results to slope zoning unit numbers, consistency between the contour line results and the corresponding slope zoning units, cross-sectional results, and slope structure traces can be ensured when subsequent deliverables are retrieved.

[0069] like Figure 4 As shown, a mapping and geographic information service deliverable package is generated based on the 3D survey model of the engineering slope and the slope structure trace results. The package includes the 3D survey model of the engineering slope, orthophotos, point cloud files, cross-section results, contour line results, slope structure trace results, supplementary survey record files, and metadata files. When generating the metadata file, the mapping and geographic information service task identifier, slope zoning unit number, first photogrammetric acquisition parameters, second photogrammetric acquisition parameters, supplementary survey trigger conditions, boundaries of continuous areas, zoning local update records, and deliverable file index are written into the metadata file. The supplementary survey record file records the slope zoning unit number where supplementary surveys occurred, the continuous area formed by supplementary survey candidate points, the second photogrammetric acquisition parameters, the second multi-view image data acquisition time, and the zoning local update results.

[0070] After generating the surveying and mapping geographic information service deliverable package, a zoning consistency verification is performed on the package. Zoning consistency verification includes verifying the spatial positional deviation of corresponding slope zoning units based on independent checkpoints, and verifying the boundary stitching deviation of adjacent slope zoning units based on zoning boundary constraint points. The spatial positional deviation and boundary stitching deviation are then written into the supplementary survey record file. Independent checkpoints are not involved in bundle adjustment, therefore they can be used to independently evaluate the spatial positional accuracy of the 3D survey model of the engineering slope. Zoning boundary constraint points are located at the intersection of adjacent slope zoning units, therefore they can be used to evaluate the point cloud boundary stitching quality after local zoning updates.

[0071] In an example of accuracy verification calculation, suppose the measured plane coordinates of an independent checkpoint are 3250.12 meters east and 1840.35 meters north, and the plane coordinates of the corresponding point in the 3D survey model of the engineering slope are 3250.15 meters east and 1840.37 meters north. Then, what is the plane position deviation of this independent checkpoint? Calculate according to the following standard mathematical formula: Among them, planar position deviation This represents the deviation of the corresponding point in the 3D survey model of the engineering slope from the measured plane position of the independent checkpoint, expressed in meters. The calculated result is approximately 0.036 meters, or 3.6 centimeters.

[0072] If the measured elevation of this independent checkpoint is 126.45 meters, and the elevation of the corresponding point in the 3D survey model of the engineering slope is 126.49 meters, then the elevation deviation of this independent checkpoint is... Calculate according to the following standard mathematical formula: Among them, elevation deviation This represents the deviation of the corresponding point in the 3D survey model of the engineering slope from the measured elevation of the independent checkpoint, in meters. The calculated result above is 0.04 meters, or 4 centimeters.

[0073] If the updated point cloud boundary elevations of two adjacent slope partition units at the same partition boundary constraint point are 118.32 meters and 118.36 meters respectively, then the boundary splicing deviation is... Calculate according to the following standard mathematical formula: Among them, boundary splicing deviation This represents the spatial deviation at the boundary splicing position of adjacent slope zoning units, expressed in meters. The calculated result above is 4 centimeters. If the preset allowable boundary splicing deviation is 5 centimeters, then the boundary splicing result of the adjacent slope zoning units meets the zoning consistency verification requirements.

[0074] To illustrate the improvement of this embodiment compared to existing slope 3D reconstruction methods, a comparative experiment was conducted in the same engineering slope test area. The test area was a Class III rock slope with a horizontal length of approximately 200 meters and a maximum elevation difference of approximately 65 meters. The slope surface included two plateau areas, several concave areas at the slope toe, and some areas obscured by vegetation. The comparative method employed a single-use UAV multi-view image acquisition process, feature matching, bundle adjustment, dense point cloud generation, and 3D model reconstruction, without performing the identification of supplementary measurement candidate points based on point cloud measurability data and the generation of second photogrammetric acquisition parameters. Under the same initial acquisition conditions, this embodiment further performs slope zoning unit establishment, point cloud measurability data generation, continuous area identification of supplementary measurement candidate points, generation of second photogrammetric acquisition parameters, acquisition of second multi-view image data, and local zoning updates.

[0075] Table 1 Comparison of 3D Slope Reconstruction Methods Point cloud integrity can be calculated using the following standard mathematical formula: Among them, point cloud completeness This indicates the proportion of the effective point cloud coverage area of ​​the engineering slope to the area that the engineering slope should cover; the area that the engineering slope should cover. This indicates the area within the engineering slope survey area that should form a valid point cloud; the area of ​​missing point cloud measurements. This indicates the area where no effective point cloud was formed due to occlusion, weak texture, or insufficient viewpoint.

[0076] In a specific example, the area that the engineering slope should cover For example, the missing area in the point cloud corresponding to the publicly disclosed 3D slope reconstruction method is 30,000 square meters. If the area is 2310 square meters, then the corresponding point cloud completeness is... for: The missing point cloud area corresponding to this implementation method If the area is 570 square meters, then the point cloud completeness rate corresponding to this implementation method is... for: As can be seen, this embodiment identifies continuous regions of candidate points for supplementary measurement and generates second photogrammetric acquisition parameters, enabling the supplementary tilt view image and the supplementary slope foot lateral view image to be acquired for areas with low measurability. After local updates in the partitioned areas, the point cloud integrity rate in this test area is increased from about 92.3% to about 98.1%.

[0077] like Figure 5 As shown, the point cloud integrity rates were approximately 92.3%, 96.7%, 98.1%, and 98.0% during the initial reconstruction, first supplementary measurement update, second supplementary measurement update, and result package verification stages, respectively. This curve reflects that in this experimental area, after continuous area identification and directional supplementary measurement of candidate points, the integrity of the dense point cloud data of the engineering slope shows an improving trend. Figure 6 As shown, the mean square error of the independent checkpoint plane position, the mean square error of the elevation, and the maximum deviation of the partition boundary splicing are all lower than those of the existing slope three-dimensional reconstruction methods. This indicates that in this test area, the independent checkpoint verification and partition boundary constraint point splicing in this embodiment are beneficial to improving the accuracy of the results and the consistency of the partition splicing.

[0078] The above experimental data are used to illustrate the implementation of this method in an engineering slope test area, and are not intended to limit the specific range of values ​​for engineering slope type, number of acquired images, number of supplementary images, and point cloud integrity rate. In practical applications, the experimental data may vary depending on the scale of the engineering slope, the complexity of the terrain, the parameters of the acquisition equipment, the mapping scale, and the requirements of surveying and mapping geographic information service results.

[0079] In another embodiment, if the engineering slope is a soil slope with significant vegetation obstruction, the obstruction risk level can be set to three levels: high, medium, and low. Furthermore, slope zoning units with higher obstruction risk levels can have a higher proportion of toe lateral view images in the first photogrammetric acquisition parameters. In this case, in the point cloud measurability data, areas with insufficient texture support and obvious local point cloud discontinuities can be prioritized as candidate points for supplementary measurement. When generating the second photogrammetric acquisition parameters, the acquisition ratio of supplementary toe lateral view images can be increased to enhance the point cloud continuity between the vegetation obstruction edge area and the toe area.

[0080] In another embodiment, if the engineering slope is a steep rock slope, the candidate directions of the structural traces can be referenced more extensively from the linear texture directions and plateau line directions in existing slope image data. Multiple camera tilt acquisition sub-parameters with different camera orientations can be set in the first photogrammetric acquisition parameters. For areas with high spatial density of updated slope structural trace points, more key cross-sectional locations can be generated, and fissure traces, slope erosion traces, and step boundary traces can be written into different layers of the slope structural trace results, thereby forming a more suitable mapping and geographic information service package for the analysis of steep rock slope engineering.

[0081] As can be seen from the above specific implementation methods, this invention is based on slope zoning units and acquisition status records. It generates measurable point cloud data from initial camera pose data, initial dense point cloud data of the engineering slope, and acquisition status records. The measurable point cloud data is used to identify candidate points for supplementary measurement and their continuous regions. Then, based on the continuous regions and corresponding initial camera pose data, second photogrammetric acquisition parameters are generated. Finally, the initial dense point cloud data of the engineering slope is locally updated by zoning using second multi-view image data. This technical process can perform targeted supplementary measurement of missing point cloud areas in engineering slopes caused by occlusion, weak texture, and insufficient viewing angle. It also establishes a unified association between the supplementary measurement process and the models, images, point clouds, cross-sections, contour lines, slope structure traces, supplementary measurement record files, and metadata files in the surveying and mapping geographic information service deliverables package. This provides an implementable technical solution for 3D surveying and mapping geographic information services of engineering slopes.

[0082] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A three-dimensional surveying method for engineering slopes based on photogrammetry, characterized in that, Includes the following steps: S1. Obtain engineering slope survey task information and establish zoning benchmark data for the engineering slope survey area. The engineering slope survey task information includes surveying and mapping geographic information service task identifiers. The zoning benchmark data includes slope top line, slope toe line, platform line, obstruction boundary, existing ground constraint points, and existing slope image data. S2. Based on the slope crest line, slope toe line, platform line, and occlusion boundary, the engineering slope survey area is divided into several slope partition units, and a data acquisition status record is established for each slope partition unit. The data acquisition status record includes the slope partition unit number, target ground resolution, candidate direction of structural trace, occlusion risk level, and supplementary measurement trigger conditions. S3. Generate first photogrammetric acquisition parameters according to the acquisition status record, and acquire first multi-view image data and ground constraint point data of the engineering slope survey area according to the first photogrammetric acquisition parameters. The first multi-view image data includes orthophoto image, oblique image and slope toe lateral image. S4. Perform image filtering, corresponding point matching and bundle adjustment on the first multi-view image data to generate initial camera pose data, initial sparse point cloud data and initial engineering slope dense point cloud data. S5. Based on the initial camera pose data, the initial engineering slope dense point cloud data, and the acquisition status record, generate point cloud measurability data for each slope partition unit. The point cloud measurability data includes the number of times the image is visible, the range of ray intersection angles, the local point cloud discontinuity boundaries, and the texture support status. S6. Based on the measurability data of the point cloud, classify the initial dense point cloud data of the engineering slope into initial surface points, initial slope structure trace points, and supplementary measurement candidate points; when the supplementary measurement candidate points within the same slope partition unit form a continuous area and meet the supplementary measurement triggering condition, generate second photogrammetric acquisition parameters based on the location, boundary, slope partition unit number, and initial camera pose data corresponding to the continuous area; S7. Acquire second multi-view image data according to the second photogrammetric acquisition parameters. The second multi-view image data includes at least one of supplementary oblique view image and supplementary slope toe lateral view image. Incorporate the second multi-view image data into the corresponding slope partition unit. Perform partitioned local updates on the initial engineering slope dense point cloud data to obtain updated engineering slope dense point cloud data. Reclassify the updated engineering slope dense point cloud data to obtain updated surface points and updated slope structure trace points. S8. Construct a three-dimensional survey model of the engineering slope based on the updated surface points, generate slope structure trace results based on the updated slope structure trace points, and generate a surveying and mapping geographic information service result package based on the three-dimensional survey model of the engineering slope and the slope structure trace results. The surveying and mapping geographic information service result package includes the three-dimensional survey model of the engineering slope, orthophoto, point cloud file, cross-section results, contour line results, slope structure trace results, supplementary survey record file, and metadata file.

2. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 1, characterized in that, In step S2, when dividing the slope into partition units, the longitudinal division of the slope is first determined according to the slope crest line, the slope toe line and the platform line, and the transverse division of the slope is determined according to the shading boundary and the existing ground constraint points. The longitudinal division of the slope and the transverse division of the slope are then superimposed to form the slope partition unit.

3. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 1, characterized in that, In step S2, the candidate direction of the structural trace is determined based on the linear texture direction of the existing slope image data within the slope partition unit, the direction of the platform line adjacent to the slope partition unit, and the direction of the slope toe line adjacent to the slope partition unit. When the angle between the linear texture direction of the existing slope image data within the slope partition unit and the direction of the platform line adjacent to the slope partition unit or the direction of the slope toe line adjacent to the slope partition unit is within a preset angle range, the linear texture direction is written into the acquisition status record.

4. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 3, characterized in that, In step S3, when generating the first photogrammetric acquisition parameters, orthophoto acquisition sub-parameters, tilt acquisition sub-parameters, and slope toe lateral acquisition sub-parameters are set for each slope partition unit. The tilt acquisition sub-parameter sets the camera orientation based on the candidate direction of the structural trace, and the slope foot lateral acquisition sub-parameter sets the acquisition distance and camera pitch angle based on the occlusion risk level.

5. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 1, characterized in that, In step S3, the ground constraint point data includes partition boundary constraint points, slope transition constraint points, and independent check points; The partition boundary constraint points are located at the junction of adjacent slope partition units, the slope transition constraint points are located at the top line, the platform line and the toe line, and the independent checkpoints do not participate in the bundle adjustment.

6. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 1, characterized in that, In step S3, when acquiring the first multi-view image data, the image view type, slope partition unit number, camera orientation, acquisition height, acquisition distance and acquisition time are written into the image metadata of the corresponding image. The image view types include orthographic view type, oblique view type, and slope foot lateral view type.

7. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 6, characterized in that, In step S4, when performing the matching of corresponding points, a partitioned image association table is established based on the image metadata, and within the same slope partition unit, corresponding points between images of different types in the orthophoto view type, the tilted view type, and the slope foot lateral view type are matched preferentially. When the number of points with the same name within the same slope partition unit is less than a preset number, the slope partition unit is added to the list of partitions to be evaluated.

8. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 1, characterized in that, In step S5, when generating the point cloud measurability data, the correspondence between the image rays and the point cloud points in the initial engineering slope dense point cloud data is established based on the initial camera pose data, and the number of times the image is visible and the range of the ray intersection angle corresponding to each point cloud point are counted. When the number of times the image is visible is lower than the preset number of times it is visible, or when the range of the ray intersection angle does not meet the preset intersection angle range, the corresponding point cloud point is marked as a low measurability point.

9. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 8, characterized in that, In step S6, when classifying the initial engineering slope dense point cloud data into the initial surface points, the initial slope structure trace points, and the supplementary measurement candidate points, the following classification rules are adopted: Point cloud points that are spatially continuous and whose local flatness meets the preset flatness conditions are determined as the initial surface points; The point cloud points that satisfy the linear texture condition, have normal variation boundaries, and cross-image visibility in the texture support state are determined as the initial slope structure trace points. Points in the point cloud that belong to the low measurability points and are located near the discontinuity boundary of the local point cloud are identified as candidate points for supplementary measurement.

10. The three-dimensional surveying method for engineering slopes based on photogrammetry according to claim 1, characterized in that, In step S6, when generating the second photogrammetric acquisition parameters, the supplementary measurement center is determined based on the location of the continuous area, the supplementary measurement coverage is determined based on the boundary of the continuous area, the existing acquisition blind direction is determined based on the initial camera pose data corresponding to the continuous area, and the acquisition direction that avoids the existing acquisition blind direction is determined as the supplementary measurement acquisition direction.

Images