A geological area mapping precision optimization method

Abstract

The present application relates to the technical field of topographic mapping, and particularly relates to a geological area mapping precision optimization method, which comprises the following steps: obtaining initial mapping data of a target geological area and extracting candidate boundary segments to form a candidate boundary belt; generating a conflict belt priority sequence according to a cross-scale stability index of the candidate boundary belt, a difference in neighborhood feature distribution on both sides and a misjudgment influence value; performing normal penetration re-measurement and tangential tracking re-measurement according to the conflict belt priority sequence, determining a boundary position and updating the candidate boundary belt; and performing constraint reconstruction according to the updated candidate boundary belt and outputting a mapping result. The present application can improve the accuracy of key boundary identification, enhance the re-measurement pertinence and maintain the real boundary in the reconstruction result.

Description

Technical Field

[0001] This invention relates to the field of topographic mapping technology, and specifically to a method for optimizing the accuracy of geological area mapping. Background Technology

[0002] The location of surface boundaries, slope breaks, zoning transition zones, and local topographic reconstruction results directly affect the reliability of basic surveys, engineering planning, disaster identification, slope treatment, and earthwork calculations. Existing surveying and mapping methods mostly rely on one-time measurements, fixed threshold determination, and unified reconstruction. Typically, suspected boundaries are extracted based on initial data, and then the mapping results are generated through local supplementary measurements or overall smoothing. In areas with significant topographic relief, frequent changes in surface material, numerous obstructions or missing measurements, or overlapping boundaries, this approach is prone to problems such as discrete boundary segments, incorrect boundary connections, and the mixing of important and noisy boundaries. Especially when resurveying resources are limited, existing technologies often struggle to accurately determine which boundary segments should be prioritized for resurveying, resulting in a lack of targeted resurveying order. Furthermore, boundaries are easily weakened by unified smoothing during subsequent reconstruction, leading to inconsistencies between the regional division results and the actual topographic boundaries, thus affecting the credibility and subsequent usability of the surveying and mapping results.

[0003] Therefore, accurately selecting key boundaries based on initial data, rationally arranging the retesting order, and maintaining true boundaries during the reconstruction process have become key areas for continuous improvement of related technologies. Summary of the Invention

[0004] This invention provides a method for optimizing the accuracy of geological regional surveying, which addresses at least the problems of accurately determining key boundaries in complex surveying data, rationally implementing priority resurveys, and maintaining true boundaries to complete terrain reconstruction.

[0005] This invention provides a method for optimizing the accuracy of geological area surveying, the method comprising: Acquire initial mapping data of the target geological area and extract candidate boundary segments based on the initial mapping data; Candidate boundary bands are formed based on candidate boundary segments, and conflict zone priority sequences are generated based on the cross-scale stability index of the candidate boundary bands, the difference in the distribution of features of the two neighboring areas, and the misjudgment impact value. The cross-scale stability index is used to indicate the stability of the candidate boundary bands at multiple scales, and the misjudgment impact value is used to indicate the degree of impact of misjudgment of the candidate boundary bands on regional division and boundary reconstruction. Perform normal crossing retest and tangential tracing retest according to the conflict zone priority sequence to determine the boundary position and update the candidate boundary zone; Based on the updated candidate boundary zones, constrained reconstruction is performed, and the mapping results of the target geological area are output.

[0006] In one possible implementation, candidate boundary segments are extracted based on initial mapping data, including: determining elevation change information, surface feature information, and observation integrity information based on the initial mapping data; and extracting candidate boundary segments based on the elevation change information, surface feature information, and observation integrity information.

[0007] In one possible implementation, forming a candidate boundary zone based on candidate boundary segments includes: splicing candidate boundary segments based on their spatial proximity and directional continuity; and determining the splicing result as a candidate boundary zone if at least one of the elevation change information, surface feature information, and observation integrity information corresponding to adjacent candidate boundary segments satisfies the same trend.

[0008] In one possible implementation, the determination of cross-scale stability indices includes: constructing boundary saliency maps corresponding to candidate boundary zones based on initial mapping data at multiple scales; extracting linear boundary skeletons based on each boundary saliency map; and determining cross-scale stability indices based on the duration, directional consistency, and morphological preservation of the linear boundary skeletons at different scales.

[0009] In one possible implementation, determining the difference in feature distribution between the two neighboring areas includes: determining a first neighboring area and a second neighboring area along the normal of the candidate boundary zone; determining geometric features in the first and second neighboring areas based on elevation change information; determining surface features in the first and second neighboring areas based on surface feature information; determining observation integrity features in the first and second neighboring areas based on observation integrity information; and determining the difference in feature distribution between the two neighboring areas based on the statistical distribution differences in geometric features, surface features, and observation integrity features between the first and second neighboring areas. Geometric features include elevation residuals, slope, and roughness; surface features include reflection intensity and texture features; and observation integrity features include point density, occlusion rate, and porosity.

[0010] In one possible implementation, the determination of the misjudgment impact value includes: determining the misjudgment impact value based on the regional elevation difference, boundary length, area difference between adjacent regions, and boundary connection relationship corresponding to the candidate boundary zone.

[0011] In one possible implementation, generating a conflict zone priority sequence includes: determining the retest priority corresponding to the candidate boundary zone based on cross-scale stability index, differences in feature distribution between the two neighboring areas, misjudgment impact value, and observation sufficiency; sorting the candidate boundary zones according to the retest priority to generate a conflict zone priority sequence; wherein, observation sufficiency is used to characterize the degree of coverage of the candidate boundary zone by the current mapping data.

[0012] In one possible implementation, performing a normal crossing retest includes: for a target conflict zone in the conflict zone priority sequence, laying out crossing survey lines along the normal of the target conflict zone; obtaining retest data corresponding to each crossing survey line; determining boundary change points based on the retest data, and determining the boundary position based on the differences in local feature distribution on both sides of the boundary change points; and updating the candidate boundary zone based on the boundary position.

[0013] In one possible implementation, performing tangential tracing retesting includes: deploying multiple overlapping tracing segments along the extension direction of the target conflict zone in the conflict zone priority sequence; acquiring retesting data corresponding to each tracing segment; determining the continuous state, termination state, or segmented state of the candidate boundary zone based on the retesting data corresponding to each tracing segment; and updating the candidate boundary zone based on the continuous state, termination state, or segmented state of the candidate boundary zone.

[0014] In one possible implementation, constraint reconstruction is performed based on the updated candidate boundary zone, including: determining the frozen boundary based on the changes in the boundary position and the changes in the boundary connectivity of the updated candidate boundary zone; using the frozen boundary as a boundary constraint, performing local terrain surface reconstruction on the area enclosed by the frozen boundary; and determining the corresponding area as an area to be confirmed if the updated candidate boundary zone does not form a frozen boundary.

[0015] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows: By extracting candidate boundary segments and forming candidate boundary bands, discrete boundary information is transformed into continuous boundary objects, thus avoiding the direct participation of local outliers in the overall reconstruction.

[0016] By introducing a cross-scale stability index, the system can determine the continuous performance of the boundary at different scales, thereby reducing the interference of single-scale noise on the boundary recognition results.

[0017] By analyzing the differences in the distribution of features in the neighborhoods on both sides of the candidate boundary zone, the system can identify the degree of separation between the terrain, surface and observation conditions on both sides of the boundary, so that boundary determination no longer depends on single-point mutations.

[0018] By determining the impact value of misjudgment, the degree of propagation impact after boundary misidentification is quantified, allowing retesting resources to be prioritized for objects that have a greater impact on regional division and boundary reconstruction.

[0019] By generating a priority sequence of conflict bands, the retesting order can be sorted and hierarchically classified, reducing invalid retesting.

[0020] By performing normal cross-traverse retesting and tangential tracing retesting, the synergistic effect of boundary position convergence and boundary continuity confirmation was achieved.

[0021] By performing constraint reconstruction based on the updated candidate boundary bands, local surface generation under boundary preservation conditions is achieved, thereby reducing the problem of the real boundary being smoothed and weakened during the reconstruction process. Attached Figure Description

[0022] Figure 1 This is a schematic flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the spatial distribution of candidate boundary segments and conflict zones in a specific embodiment of the present invention; Figure 3 This is a statistical chart of the priority of candidate boundary zone retesting in a specific embodiment of the present invention; Figure 4 This is a comparison diagram of boundary deviations before and after the re-measurement of the target conflict zone normal crossing in a specific embodiment of the present invention; Figure 5 This is a comparison of the local terrain reconstruction profile before and after freezing boundary constraints in a specific embodiment of the present invention. Detailed Implementation

[0023] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0024] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0025] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0026] Mapping accuracy optimization typically refers to improving the ability of mapping results to accurately represent the true landform by further identifying, filtering, and correcting boundary information, local topographic information, and data coverage status based on existing mapping data. Its focus is not only on the accuracy of single-point elevations or single sampling results, but also on the consistency and usability of boundary locations, regional demarcation relationships, and topographic reconstruction results across the overall space. With the continuous enrichment of multi-source mapping data acquisition methods, identifying key boundaries with greater impact on the results from the initial data and implementing targeted processing around these key boundaries has become an important technical direction for the refined expression of mapping results. Based on this, this invention proposes a mapping accuracy optimization method for geological regions.

[0027] like Figure 1 As shown, a method for optimizing the accuracy of geological area mapping includes: Acquire initial mapping data of the target geological area and extract candidate boundary segments based on the initial mapping data; After acquiring the initial mapping data for the target geological area, data from different sources are first uniformly registered to form a 3D surface dataset under the same coordinate datum. The initial mapping data can include point cloud data, image data, and control point data. After registration, outliers, duplicates, and obviously distorted sampling points in the point cloud are removed, and shadowed areas, areas with excessive reflection, and missing measurements in the image are marked. Subsequently, the target geological area is divided into grids for statistical analysis, generating elevation change information, surface feature information, and observation integrity information. Candidate boundary segments are then extracted based on the combined changes of these three types of information. These candidate boundary segments characterize the location range of potential topographic boundaries, material boundaries, or observation faults, and serve as input for subsequent boundary stitching and re-measurement ranking.

[0028] Candidate boundary segments are extracted based on the initial mapping data, including: determining elevation change information, surface feature information, and observation integrity information based on the initial mapping data; and extracting candidate boundary segments based on the elevation change information, surface feature information, and observation integrity information.

[0029] In one embodiment, to ensure that candidate boundary segments accurately reflect boundary changes rather than local noise, this section further defines the processing order of the initial mapping data, the generation method of the three types of information, and the extraction rules for candidate boundary segments. The initial mapping data preferably includes surface point clouds formed by laser scanning, image data covering the target geological area, and a small amount of control point data. Control point data is used to unify coordinates between different acquisition batches, point cloud data is used to characterize surface elevation and topographic relief, and image data is used to characterize surface texture, color, and reflectance changes.

[0030] After coordinate unification, outlier removal and density balancing are performed on the point cloud data, followed by distortion correction and occlusion marking on the image data. Outlier removal can be achieved using a neighborhood distance check; if the average distance between a sampling point and its neighboring sampling points is significantly greater than the local average, the sampling point is identified as an outlier and deleted. Density balancing is used to avoid bias in subsequent statistical results caused by excessively dense point clouds in certain areas; representative points are typically retained using a preset grid. The grid spacing can be set according to the average point spacing and target accuracy requirements. When the average point spacing is small, the grid spacing can be set to 0.2m to 0.5m; when the average point spacing is large, the grid spacing can be set to 2 to 3 times the average point spacing. Elevation variation information reflects the local surface undulation and can be obtained by statistically analyzing the elevation range within the grid, elevation jumps between adjacent grids, and local slope inflection trends.

[0031] When elevation changes continuously increase or decrease within adjacent areas, it usually indicates a significant topographic transition. Surface feature information reflects the surface material and texture, and can be determined based on changes in image grayscale, color distribution, and point cloud reflectance intensity. When there are boundaries between bare soil and vegetation, rock and backfilled areas, or road edges, surface feature information will show continuous abrupt changes. Observation integrity information reflects the coverage of the current mapping data and can be determined based on point density within a unit grid, image occlusion ratio, and missing measurement range. When the point density in a certain area is significantly lower than the surrounding area, or when there is continuous shadow occlusion in the image, the observation integrity information of that area is marked as abnormal. The extraction of candidate boundary segments is based on the joint judgment results of these three types of information.

[0032] In specific processing, grids with elevation changes reaching the preset change level are first selected, followed by grids with continuous abrupt changes in surface feature information. Simultaneously, areas with abnormal observation integrity information but insignificant elevation changes and surface features are suppressed to reduce false boundaries caused by occlusion and missing measurements. The preset change level can be set according to the degree of undulation in the survey area; a lower threshold is preferred in areas with relatively flat terrain, while a higher threshold is preferred in areas with significant terrain undulation.

[0033] After screening, spatially adjacent candidate locations with the same main direction are connected to form candidate boundary segments. The main direction can be determined based on the extension direction of adjacent sampling points or the local fitted centerline. When the distance between adjacent candidate locations is no greater than twice the current grid spacing, and the directional deviation does not exceed 15° to 30°, they can be considered to belong to the same candidate boundary segment. For candidate boundary segments that are too short and cannot form a stable extension trend, they can be directly deleted or merged into the nearest longer segment with the same trend. The minimum retention length is usually no less than three times the current grid spacing. Through the above processing, the resulting candidate boundary segments retain the main location of the true boundary while suppressing the interference of local noise, short-term abrupt changes, and missing observations on subsequent processing.

[0034] Candidate boundary bands are formed based on candidate boundary segments, and conflict zone priority sequences are generated based on the cross-scale stability index of the candidate boundary bands, the difference in the distribution of features of the two neighboring areas, and the misjudgment impact value. The cross-scale stability index is used to indicate the stability of the candidate boundary bands at multiple scales, and the misjudgment impact value is used to indicate the degree of impact of misjudgment of the candidate boundary bands on regional division and boundary reconstruction. After extracting candidate boundary segments, these segments are organized and screened to form candidate boundary zones that can participate in subsequent retesting and ranking. Once formed, cross-scale stability indices, differences in the distribution of features between the two sides' neighborhoods, and the impact of misjudgments are determined. Combined with the coverage of each candidate boundary zone by the current mapping data, a conflict zone priority sequence is generated. This sequence indicates the boundary areas that need to be prioritized for subsequent retesting, avoiding the application of the same retesting strategy to all boundary objects. This ensures that retesting resources are prioritized for areas with high stability, clear separation between the two sides, and significant impact from misjudgments.

[0035] Forming candidate boundary zones based on candidate boundary segments includes: splicing candidate boundary segments based on their spatial proximity and directional continuity; and determining the splicing result as a candidate boundary zone if at least one of the elevation change information, surface feature information, and observation integrity information corresponding to adjacent candidate boundary segments satisfies the same trend.

[0036] In one embodiment, the splicing process of candidate boundary segments is preferably performed in the order of "nearest neighbor first, then direction, then trend" to ensure that the formed candidate boundary band has both spatial continuity and boundary significance. First, a segment index table is established for all candidate boundary segments, and the start coordinates, end coordinates, center position, local main direction, and segment length of each candidate boundary segment are recorded. Then, adjacent segments within a preset distance range are searched one by one for each candidate boundary segment.

[0037] The preset distance range can be set based on the current grid spacing or the average point spacing, typically 2 to 3 times the current grid spacing, or 3 to 5 times the average point spacing. If the endpoint distance between adjacent segments exceeds this range, they will not participate in this round of stitching. For segment pairs that meet the distance conditions, the directional continuity relationship is then determined.

[0038] The directional continuity is determined by the angle between the main directions of the segments; the smaller the angle, the more consistent the boundary extension trend. Generally, the directional deviation threshold is set between 15° and 30°. When the terrain boundary is relatively straight, a smaller threshold is preferred; when the boundary has natural bends or the mapping data contains slight noise, a larger threshold is preferred.

[0039] After filtering by distance and direction, the elevation change information, surface feature information, and observation integrity information corresponding to adjacent candidate boundary segments are checked to see if there is a consistent trend. A consistent trend does not require all three types of information to be completely identical simultaneously; rather, it requires that at least one type of information shows continuous enhancement, continuous weakening, or stable continuation in the segment connection direction. For example, if adjacent segments both show increased elevation transitions, or both show continuous abrupt changes in reflectance intensity, or both show continuous sparse observation zones, then the trend can be considered consistent.

[0040] If the distance, direction, and trend conditions are all met simultaneously, the corresponding segments will be spliced ​​to form new boundary objects. Splicing can be done using endpoint connections or centerline smoothing. Segments that are significantly shorter than average and located between two longer segments can be incorporated into the splicing result as transitional segments. After splicing, a foldback check and a self-intersection check are performed. If the splicing result shows obvious foldback, closure anomalies, or abrupt changes in boundary direction, the splicing is cancelled. Objects that are too short and fail to form stable connections with other segments can be retained as pending segments or have their priority reduced in subsequent sorting. Through this process, scattered boundary segments can be integrated into candidate boundary bands that can express continuous boundary trends, providing stable objects for subsequent multi-scale judgment and retesting sorting.

[0041] The determination of cross-scale stability indices includes: constructing boundary saliency maps corresponding to candidate boundary zones based on initial mapping data at multiple scales; extracting linear boundary skeletons based on each boundary saliency map; and determining cross-scale stability indices based on the duration, directional consistency, and morphological preservation of the linear boundary skeletons at different scales.

[0042] In one embodiment, the cross-scale stability index is determined by the significant performance of the boundaries at multiple scales to avoid accidental mutations at a single scale being mistaken for stable boundaries.

[0043] In practice, firstly, boundary saliency maps at multiple scales are constructed based on the initial mapping data. These boundary saliency maps represent the spatial distribution of boundary response intensity. When constructing these maps, geometric saliency responses are extracted from elevation change information, and surface saliency responses are extracted from surface feature information. Then, observation integrity information is combined to suppress missing and occluded areas, ultimately forming boundary saliency maps at each scale. The number of scales is typically three to five. If the survey area is small and the point cloud is dense, four scales—0.5m, 1m, 2m, and 4m—can be used; if the survey area is large or the sampling is sparse, the overall scale range can be increased.

[0044] After the boundary saliency map is formed, a linear boundary skeleton is extracted from the boundary saliency map at each scale. The linear boundary skeleton is used to represent the center direction and continuous structure of the salient boundary region, and can be obtained by thinning, centerline extraction, or connected region compression. Then, the skeleton results corresponding to the same candidate boundary band at different scales are matched. Three main aspects are considered during matching: the first is the continuity length, i.e., the effective length of the candidate boundary band that can be identified at multiple scales; the second is the degree of directional consistency, i.e., whether the deviation of the main direction of the skeleton at different scales remains within a preset range; and the third is the degree of shape preservation, i.e., whether the bending positions, extension trends, and overall contour remain stable. If a candidate boundary band is salient only at one scale and cannot form a continuous skeleton at other scales, its stability is low. If a candidate boundary band has a continuous skeleton at multiple scales, and the directional changes are small and the contour position offset is limited, its stability is high.

[0045] To facilitate subsequent ranking, the cross-scale stability index can be divided into three levels: high, medium, and low, or five levels. The ranking criteria are typically determined by the proportion of the continuous length to the total length of the candidate boundary zone, the upper limit of directional deviation, and the degree of contour overlap. When the terrain boundary is clear and surface interference is minimal, a stricter ranking rule can be adopted; when the terrain is complex and there is significant local occlusion, the matching conditions can be appropriately relaxed. Through multi-scale judgment, true continuous boundaries can be distinguished from local noise boundaries, texture pseudo-boundaries, and single-observation anomalies, allowing subsequent priority ranking to be based on more stable boundary objects.

[0046] The determination of the differences in feature distribution between the two neighboring areas includes: determining the first and second neighboring areas along the normal of the candidate boundary zone; determining the geometric features in the first and second neighboring areas based on elevation change information; determining the surface features in the first and second neighboring areas based on surface feature information; determining the observation integrity features in the first and second neighboring areas based on observation integrity information; and determining the differences in feature distribution between the two neighboring areas based on the statistical distribution differences in geometric features, surface features, and observation integrity features between the first and second neighboring areas. Among these, geometric features include elevation residuals, slope, and roughness; surface features include reflection intensity and texture features; and observation integrity features include point density, occlusion rate, and porosity.

[0047] In one embodiment, the difference in the distribution of features between the two neighboring areas is determined by statistical comparison of the normal bilateral neighborhoods to determine whether there is persistent geometric separation, surface separation, or observation state separation on both sides of the candidate boundary zone. Specifically, the local normal direction of the candidate boundary zone is first determined. The normal direction can be obtained by rotating the tangential direction of the boundary zone centerline by 90°, or it can be determined based on the normal direction of the local fitted curve.

[0048] After the normal direction is determined, a first neighborhood and a second neighborhood are constructed on both sides of the candidate boundary band. The neighborhood width should be set according to the current grid spacing, average point spacing, and boundary complexity. Generally, the neighborhood width is 2 to 5 times the current grid spacing, or 10 to 20 times the average point spacing. The neighborhood length can be adaptively adjusted according to the local length of the candidate boundary band to cover the main changing areas on both sides of the boundary.

[0049] After the neighborhood is determined, geometric features are extracted first based on elevation change information, then surface features are extracted based on surface feature information, and finally observation integrity features are extracted based on observation integrity information. Geometric features include elevation residuals, slope, and roughness. Elevation residuals characterize the degree of deviation of local elevation from the reference surface, slope characterizes surface tilt changes, and roughness characterizes the dispersion of surface undulations. Surface features include reflectance intensity and texture features. Reflectance intensity characterizes the differences in response of different materials to the mapping signal, and texture features characterize grayscale changes, color changes, and local texture density. Observation integrity features include point density, occlusion rate, and porosity. Point density reflects the density of local sampling coverage, occlusion rate reflects the proportion of observations that are obscured, and porosity reflects the proportion of continuously missing areas.

[0050] After feature extraction, the mean level, dispersion, distribution range, and trend of the two neighboring regions are statistically analyzed. If there are sustained significant differences in geometric features between the two neighboring regions, it usually indicates that the boundary has a strong topographical separation effect. If there are sustained differences in surface features but no significant differences in geometric features, it may correspond to a boundary with changes in surface material. If the main differences are concentrated in the observation integrity features, it may correspond to an occlusion boundary or a sampling break boundary. To reduce the interference of single outliers, a local window sliding statistical method is preferred to calculate the boundary zone segment by segment, and then the statistical results of each segment are summarized. If most segments of the boundary zone show stable differences, the feature distribution differences between the two neighboring regions are considered significant. If only a few segments show isolated differences, the separation level of the candidate boundary zone can be reduced. Through this two-sided neighbor statistical comparison method, the authenticity of the boundary can be judged from the regional distribution level, rather than relying solely on the abrupt changes of individual sampling points on the boundary line.

[0051] The determination of the impact value of misjudgment includes: determining the impact value of misjudgment based on the regional elevation difference, boundary length, area difference between adjacent regions, and boundary connection relationship corresponding to the candidate boundary zone.

[0052] In one embodiment, the impact value of misjudgment is obtained through a joint determination of regional elevation difference, boundary length, area difference between adjacent regions, and boundary connectivity, to characterize the degree of impact on subsequent regional delineation and boundary reconstruction after a candidate boundary zone is misidentified. Regional elevation difference reflects the intensity of the difference in overall terrain height between the regions on both sides of the candidate boundary zone. The larger the regional elevation difference, the more likely the candidate boundary zone is to serve as a significant terrain boundary. Boundary length reflects the extent of the candidate boundary zone's extension within the survey area. The longer the boundary length, the larger the spatial range affected by misjudgment. Area difference between adjacent regions reflects the degree of imbalance in area between the regions on both sides of the boundary. When one side of the region is significantly larger than the other, the boundary often serves as a primary dividing line. Boundary connectivity reflects whether the candidate boundary zone forms an intersection, bifurcation, closure, or turning structure with other candidate boundary zones. The more complex the connectivity, the easier it is for misjudgment to propagate to multiple regions.

[0053] In practice, the four types of factors can be classified into three levels: small, medium, and large; boundary length can be classified into three levels: short, medium, and long; area difference between adjacent areas can be classified into three levels: balanced, somewhat unbalanced, and significantly unbalanced; and boundary connection relationship can be classified into three levels: independent, connected, and intersecting. After classification, the misjudgment impact value is determined comprehensively according to preset rules. If a candidate boundary zone simultaneously possesses a large regional elevation difference, a relatively long boundary length, a significant area difference, and a complex connection relationship, the misjudgment impact value is set to a high level; if the candidate boundary zone exists only within a short local area, and the differences between the two sides are small and the connection relationship is simple, the misjudgment impact value is set to a low level.

[0054] To prevent individual factors from amplifying the impact of misjudgments, constraints can be set. For example, if the boundary length is lower than the minimum length of concern, even if the regional elevation difference is large, the misjudgment impact value will only be increased by one level, instead of directly increasing it to the highest level. The minimum length of concern is usually no less than 5 times the current grid spacing. This evaluation method allows the misjudgment impact value to reflect both the importance of the boundary and the propagation impact of the misjudgment on subsequent processes, thus providing a more practical basis for prioritizing retesting.

[0055] The process of generating a conflict zone priority sequence includes: determining the retest priority of candidate boundary zones based on cross-scale stability indices, differences in feature distribution between the two neighboring areas, the impact of misjudgments, and observation sufficiency; sorting the candidate boundary zones according to the retest priority to generate a conflict zone priority sequence; wherein, observation sufficiency is used to characterize the degree of coverage of the candidate boundary zones by the current mapping data.

[0056] In one embodiment, the conflict zone priority sequence is generated by jointly ranking cross-scale stability indices, differences in feature distribution between adjacent neighborhoods, misjudgment impact values, and observation sufficiency. Observation sufficiency characterizes the degree of coverage of candidate boundary zones by the current mapping data. Observation sufficiency can be determined based on point cloud coverage density, image coverage integrity, occlusion ratio, and missing measurement ratio within the candidate boundary zone. Observation sufficiency is high when most locations of the candidate boundary zone are covered by continuous point clouds and continuous images; observation sufficiency is low when a large portion of the candidate boundary zone falls within shadow areas, hole areas, or sparsely sampled areas.

[0057] When determining the priority of retesting, the preferred order is "first determine the credibility of the boundary, then determine the consequences of misjudgment, and finally determine whether supplementary coverage is needed." Specifically, if the cross-scale stability index is high and the differences in the distribution of features in the two neighboring regions are significant, it indicates that the candidate boundary zone has a strong boundary foundation. On this basis, if the impact value of misjudgment is high, the priority of retesting should be further increased. If the observation sufficiency is low, it indicates that although the boundary foundation is strong, the current data does not support the boundary sufficiently, and the priority of retesting should also be increased.

[0058] Conversely, if the cross-scale stability index is low and the differences in the feature distributions of the two neighboring regions are not significant, then even if the observation sufficiency is low, it will not be directly included in the highest priority list, but will be reserved as an object to be further verified. In the ranking implementation, candidate boundary zones can first be divided into three levels: high priority, medium priority, and low priority. Then, within each level, they can be further subdivided according to the impact of misjudgments and the observation sufficiency. High-priority objects correspond to the conflict zones that most need priority retesting, medium-priority objects correspond to secondary conflict zones, and low-priority objects correspond to general conflict zones.

[0059] For multiple candidate boundary zones with the same priority, they can be further sorted according to boundary length, boundary connection complexity, or coverage gap ratio, so that subsequent retesting tasks can prioritize objects with greater impact. After sorting, a conflict zone priority sequence is generated according to the sorting results, and the spatial coverage of each candidate boundary zone is expanded to form the corresponding conflict zone. The expansion width can be set according to the neighborhood width, boundary complexity, and observation sufficiency. The lower the observation sufficiency and the more complex the boundary orientation, the larger the expansion width should be. Through this sorting and expansion method, limited retesting resources can be prioritized for boundary areas that truly need confirmation and whose misjudgment would have a significant impact on regional delineation and boundary reconstruction.

[0060] Perform normal crossing retest and tangential tracing retest according to the conflict zone priority sequence to determine the boundary position and update the candidate boundary zone; After establishing a priority sequence for conflict zones, retesting tasks are scheduled from highest to lowest priority. First, normal cross-sectional retesting is performed on higher-priority conflict zones to confirm the existence of stable differences on both sides of the boundary and to locate boundary changes. Then, tangential tracing retesting is performed along the same conflict zone to confirm the continuity, termination position, and segmentation relationship of the boundary along its extension direction. The results of the normal cross-sectional retesting and the tangential tracing retesting are combined to form the boundary update result, which includes at least the boundary center position, boundary extension range, boundary continuity status, and the segment to be retested. For segments with consistent retesting results, they are directly retained as updated candidate boundary zones; for segments with inconsistent retesting results, they are marked as segments to be retested and given priority in subsequent retesting rounds.

[0061] Performing normal cross-traverse retesting includes: laying out cross-traverse survey lines along the normal direction of the target conflict zone in the conflict zone priority sequence; obtaining retesting data corresponding to each cross-traverse survey line; determining boundary change points based on the retesting data, and determining the boundary position based on the differences in local feature distribution on both sides of the boundary change points; and updating the candidate boundary zone based on the boundary position.

[0062] In one embodiment, the normal crossing retest further defines the selection method of the target conflict zone, the layout method of the crossing survey line, the identification method of the boundary change point, and the update rule of the boundary position. During the retest scheduling, conflict zones with high priority are selected first, and within the same conflict zone, sections with large changes in boundary orientation, strong differences in the distribution of features of the two neighboring areas, or low observation sufficiency are selected as the first round of retest sections.

[0063] After determining the sections for the first round of resurveying, the normal direction is first determined based on the local tangential direction of the candidate boundary zone, and then the crossing survey lines are laid out along the normal direction. The spacing of the crossing survey lines should not be too large to avoid missing local boundary changes, nor should it be too small to avoid excessive duplicate sampling. Generally, the spacing of the crossing survey lines can be set to 2 to 4 times the current grid spacing, or to 5% to 10% of the local length of the candidate boundary zone. In sections with obvious boundary bends, significant occlusion, or unstable results from the previous round of resurveying, the spacing of the survey lines can be appropriately reduced.

[0064] Each survey line should ideally cover the complete neighborhood on both sides of the boundary, rather than just a narrow area near the boundary center. The survey line length is typically set to twice the width of the neighborhood on both sides to ensure that the resurvey data includes stable areas on both sides of the boundary. Resurvey data can come from supplementary point clouds, supplementary imagery, control point resurvey results, or close-range scan results. After acquiring the resurvey data, the sampling points are first sorted according to the survey line direction, and then elevation changes, reflectance changes, texture changes, and observation integrity changes are extracted separately. The identification of boundary change points does not rely on a single abrupt change value, but is determined based on the continuous local change areas along the survey line direction.

[0065] In practice, points with significant elevation changes can be identified first, followed by points showing synchronous changes in surface features. The presence of continuous observational coverage on both sides of each point can then be checked. If a point exhibits significant geometric or surface differences on both sides, and these differences occur continuously along adjacent survey lines, then that point is identified as a boundary change point. If multiple boundary change points appear along a single survey line, the point closest to the center of the candidate boundary zone and with the longest duration of difference on both sides should be prioritized. If no stable boundary change points appear along multiple survey lines, the conflict zone should not be directly deleted; instead, the section should be marked as a low-confidence section and await further confirmation through tangential tracing resurveys.

[0066] After the boundary positions are determined, the boundary change points of each survey line need to be projected onto the centerline direction of the candidate boundary zone to form a new set of boundary center positions. If the offset between adjacent boundary center positions is within the allowable range, the candidate boundary zone is updated using a smooth connection method; if the local offset is significantly greater than the allowable range, the segment is separated from the original candidate boundary zone to form a boundary sub-segment to be further confirmed. The allowable range is usually set based on the current grid spacing, average point spacing, and survey area accuracy requirements, and is generally no greater than 1 to 2 times the current grid spacing. Through the above processing, the boundary center positions can be directly corrected using the normal crossing re-measurement results, and segments with weak boundary changes, unstable boundary positions, or multiple candidate positions can be identified from the original candidate boundary zone, providing a clearer target range for subsequent tangential tracking re-measurement.

[0067] Performing tangential tracing retests includes: deploying multiple overlapping tracing segments along the extension direction of the target conflict zone in the conflict zone priority sequence; acquiring retest data corresponding to each tracing segment; determining the continuous state, termination state, or segmented state of the candidate boundary zone based on the retest data corresponding to each tracing segment; and updating the candidate boundary zone based on the continuous state, termination state, or segmented state of the candidate boundary zone.

[0068] In one embodiment, the tangential tracing retest further defines the layout of the tracing segments, the determination method for continuous and terminated states, and the update method for candidate boundary zones. The main objective of the tangential tracing retest is not to re-determine whether there are differences on both sides of the boundary, but to confirm whether the boundary continues along the extension direction, whether it is interrupted at a local location, whether there is a significant deflection, and whether it should be split into multiple boundary segments.

[0069] In practice, the local extension direction is first determined based on the updated boundary center position after normal crossing, and then multiple overlapping tracking segments are deployed along this direction. Overlapping areas should be maintained between adjacent tracking segments to allow for comparison of the continuity of boundary positions and features. The overlap ratio is generally set to 20% to 40% of the tracking segment length. If the boundary has many bends or observations are sparse, the overlap ratio should be appropriately increased.

[0070] Within each tracking segment, corresponding remeasurement data is acquired and segmented to form boundary observation records arranged along the extension direction. The criteria for determining continuity include whether the boundary position offset is continuous, whether the boundary orientation change is gradual, and whether the characteristic differences on both sides of the boundary are stable. If the boundary center position offset in adjacent tracking segments does not exceed a preset upper limit, the boundary orientation change remains continuous, and the differences on both sides of the boundary persist in most tracking segments, then the corresponding segment is determined to be continuous.

[0071] The criteria for determining the termination state include the absence of stable boundary changes in multiple consecutive tracking segments, a significant reduction in differences on both sides of the boundary, and the observation sufficiency meeting the set requirements. To avoid misjudging occluded or missing measurement areas as termination states, the location is only determined as the termination location when there are no stable boundary changes in two or more consecutive tracking segments, and the point cloud coverage and image coverage meet the minimum coverage requirements.

[0072] Segmentation is used to characterize situations where the same candidate boundary zone exhibits significant breaks, significant deflections, or abrupt changes in boundary attributes during its extension. When the boundary position offset in adjacent tracking segments continuously exceeds the allowable range, or when there is a significant abrupt change in the boundary orientation, or when the characteristic distribution on both sides of the boundary changes from being dominated by geometric differences to being dominated by surface differences, the candidate boundary zone can be divided into two or more segmented sections.

[0073] When updating candidate boundary zones, for continuous state segments, the original boundary zone is either extended or maintained; for terminated state segments, the boundary zone is truncated at the termination position; for segmented state segments, new boundary sub-zones are generated respectively, and the start and end positions and connection relationships of each boundary sub-zone are recorded. If a local segment neither satisfies the continuous state nor can its terminated state be confirmed, it is retained as a segment to be remeasured, and in the next round of remeasurement, shorter tracking segments or additional cross-directional survey lines are prioritized.

[0074] By confirming the boundary position segment by segment along the extension direction, the boundary position update result can be expanded from single-point correction to the entire boundary update result, so that the updated candidate boundary band has a more accurate position, a clearer termination range, and a clearer segmentation structure.

[0075] Based on the updated candidate boundary zones, constrained reconstruction is performed, and the mapping results of the target geological area are output.

[0076] After updating the candidate boundary zones, the updated results are reconstructed into different states, and boundaries that meet stability conditions are used as terrain reconstruction constraints. Specifically, the changes in position and connectivity of each updated boundary are first statistically analyzed across consecutive resurvey cycles. Boundaries whose position changes and connectivity tend to be stable are then selected as frozen boundaries. Subsequently, the area enclosed by the frozen boundaries is used as an independent reconstruction unit. Point cloud data, image-aided data, and control point data within the corresponding unit are used to reconstruct the local terrain surface, outputting the mapping results for the target geological area. Areas that have not yet formed frozen boundaries are not directly involved in the formal reconstruction but are marked as areas awaiting confirmation to prevent unstable boundaries from entering the final result.

[0077] The constraint reconstruction based on the updated candidate boundary zone includes: determining the frozen boundary based on the changes in the boundary position and the changes in the boundary connection relationship of the updated candidate boundary zone; using the frozen boundary as the boundary constraint, performing local terrain surface reconstruction on the area enclosed by the frozen boundary; and determining the corresponding area as the area to be confirmed if the updated candidate boundary zone does not form a frozen boundary.

[0078] In one embodiment, to ensure that the reconstruction results maintain the true boundary shape while avoiding the direct impact of unstable boundaries on the final surface representation, this section further defines the method for determining frozen boundaries, the processing order of local terrain surface reconstruction, and the rules for retaining areas to be confirmed. After the updated boundaries are formed, a boundary update record table is first established according to the boundary number to record the center position, endpoint position, extension direction, and connected objects of each boundary in adjacent retesting rounds.

[0079] The change in boundary position reflects the degree of convergence of the boundary position in consecutive retesting rounds, and can be determined by the offset of the center position and the offset of the endpoint positions of the same boundary in adjacent rounds. The change in boundary connection relationship reflects whether the connection state between the boundary and adjacent boundaries is stable, and can be determined by whether the intersection position changes, whether the connected objects change, and whether the connection order changes.

[0080] In practice, an upper limit can be set first for the change in boundary position, and then an upper limit can be set for the change in boundary connection relationship. When the change in position of the same boundary does not exceed the upper limit for both position and connection relationship in two or three consecutive rounds of remeasurement, the boundary is determined as a frozen boundary. The upper limit for position is generally set according to the current grid spacing, average point spacing, and target accuracy requirements, and is usually set to 1 to 2 times the current grid spacing; the upper limit for connection is usually determined according to the rule of "connected objects remain unchanged, and the offset of the intersection position does not exceed the upper limit for position". Boundaries with stable positions but changing connection relationships are not frozen; boundaries with stable connection relationships but continuously shifting positions are also not frozen. After the frozen boundary is determined, the area enclosed by the frozen boundary is divided into zones. When dividing, it is preferable to first identify closed boundary loops, and then identify semi-closed areas enclosed by open boundaries and survey area boundaries. Each zone is treated as a separate reconstruction unit.

[0081] When reconstructing local terrain surfaces, point cloud data within the region is prioritized as the elevation base, and then control point data is used to constrain and correct the surface elevation within the region. When the image data within the region is of high quality, image data can also be used to assist in verifying the position of surface texture edges. During the reconstruction process, frozen boundaries are used as hard constraint boundaries, and the reconstructed surface is not allowed to continuously smooth across frozen boundaries to avoid weakening the true boundaries between adjacent regions. If the point cloud coverage within a certain region is relatively complete, the terrain surface is directly generated using local triangulation or local surface fitting methods. If there are a few missing measurements within a certain region, but the frozen boundary is complete, the missing measurements are locally supplemented under the constraints of the frozen boundary. If the missing measurement range within a certain region is too large, and the surface supplementation cannot guarantee the continuity and rationality of the surfaces on both sides of the boundary, only the measured parts are retained, and a complete surface is not forcibly generated.

[0082] Areas without frozen boundaries are uniformly marked as areas awaiting confirmation. The determination of areas awaiting confirmation is based not only on whether the boundaries are frozen, but also on whether there are large-scale obstructions, continuous holes, or multiple intersecting and unresolved boundaries within the area. Areas awaiting confirmation are not included in the final mapping results in this round of output; only their spatial extent, associated boundary numbers, and the marker for subsequent re-measurement are retained for later re-measurement. Through the above processing, the final output mapping results consist of frozen boundaries, the local topographic surfaces of the areas enclosed by each frozen boundary, and the markers for areas awaiting confirmation. This ensures that the final output results are based on stable boundaries and clearly distinguishes unstable areas from confirmed areas.

[0083] In one specific embodiment, a river valley road slope treatment survey area was selected as the implementation object. The survey area has an east-west length of 120m, a north-south width of 80m, and a total area of ​​9600m². 2 The survey area contains stepped slopes, backfilled platforms, shallow ditches, and vegetation-covered zones, with surface elevations ranging from 45.2m to 58.9m. Twelve control points were established on-site. Point cloud data was acquired using UAV laser scanning, with an average point spacing of 0.18m and a total point cloud volume of 1.152 million points. Simultaneously, ground imagery data with a resolution of 0.05m was acquired. After coordinate unification, outlier removal, and marking of obstructed areas, 1.084 million valid point cloud points were retained. The survey area was then gridded with a grid spacing of 0.5m, resulting in 38,400 grid cells.

[0084] Based on the unified surveying data, elevation change information, surface feature information, and observation integrity information were first extracted. Elevation change information mainly reflects local elevation shifts, slope breaks, and gully edges; surface feature information mainly reflects changes in reflection intensity and texture; and observation integrity information mainly reflects point density, occlusion rate, and porosity. Based on the combined changes of these three types of information, a total of 214 candidate boundary segments were extracted. These were further spliced ​​according to spatial proximity, directional continuity, and consistency of change trends, resulting in 29 candidate boundary zones, with the shortest being 4.6m and the longest 21.3m. To avoid short, noisy boundaries entering subsequent processes, objects shorter than 1.5m that did not form stable connections with other segments were directly deleted.

[0085] like Figure 2 As shown, this figure is obtained by overlaying the base elevation map of the survey area, candidate boundary zones, and high-priority conflict zones in a unified coordinate system. In the figure, gray lines represent the formed candidate boundary zones, red lines represent conflict zones entering the high-priority resurvey sequence, and blue square dots represent the locations of control points. Figure 2 It can be seen that the candidate boundary zones are mainly distributed near the slope break zone, platform edge and ditch edge, while the high priority conflict zone is mainly concentrated in the boundary intersection area, the edge of the shading area and the area where elevation change and texture change coexist.

[0086] After the candidate boundary zones were formed, the cross-scale stability index, the difference in feature distribution between the two neighboring areas, and the impact of misjudgment were calculated. The cross-scale stability index was calculated at four scales: 0.5m, 1m, 2m, and 4m. A boundary saliency map was constructed for each scale, and a linear boundary skeleton was extracted and graded according to its duration, directional consistency, and morphological preservation. Among the 29 candidate boundary zones, 8 had high cross-scale stability indices, 11 had medium indices, and 10 had low indices. The difference in feature distribution between the two neighboring areas was determined by statistical comparison of the normal bilateral neighboring areas, with the neighborhood width taken as four times the current grid spacing. Geometric features included elevation residuals, slope, and roughness; surface features included reflection intensity and texture variation; and observation integrity features included point density, occlusion rate, and porosity. The impact of misjudgment was determined by grading the regional elevation difference, boundary length, area difference between adjacent areas, and boundary connectivity.

[0087] like Figure 3 As shown in the figure, the 10 candidate boundary zones with the highest priority were selected, and their cross-scale stability, degree of separation on both sides, impact of misjudgment, and final retest priority were statistically analyzed. The figure shows that B07 has the highest retest priority at 95 points, followed by B12 at 90 points and B05 at 86 points. Eight high-priority candidate boundary zones entered the first round of retesting, 11 medium-priority candidate boundary zones entered the second round of retesting preparation, and 10 low-priority candidate boundary zones were not included in the first round of retesting and were retained as objects awaiting further verification.

[0088] The B07 conflict zone is used as an example. B07 is 42m long along the boundary mileage direction. Before the re-measurement, the average boundary deviation relative to the control line was 0.903m, with a maximum deviation of 1.28m. For B07, 15 crossing survey lines were first laid out along the normal direction, with a line spacing of 3m and a single line length of 4m, covering the complete neighborhood on both sides of the boundary. Then, 6 overlapping tracking sections were laid out along the boundary extension direction, with an overlap ratio of 30%. The normal crossing re-measurement was used to locate boundary change points, and the tangential tracking re-measurement was used to identify the boundary continuity, termination, and segmentation states. The re-measurement results show that the boundary position of B07 is continuous and stable in the section from mileage 24m to 33m, gradually weakening after mileage 33m and terminating near mileage 39m.

[0089] like Figure 4 As shown, this diagram is based on the 15 normal crossing survey lines of B07, with the boundary deviations corresponding to each survey line arranged in order of boundary mileage. In the diagram, the red line represents the boundary deviation before re-surveying, the blue line represents the boundary deviation after re-surveying, and the green line represents the reference value for the control survey lines. From... Figure 4It can be seen that the deviation before the retest was generally in the range of 0.69m to 1.28m, and after the retest, it converged to the range of 0.15m to 0.34m. Calculations show that the average boundary deviation of B07 decreased from 0.903m to 0.221m, a decrease of 75.5%; the root mean square value of the boundary deviation decreased from 0.919m to 0.228m, a decrease of 75.2%. This result indicates that the boundary position can be significantly converged after the joint execution of the normal crossing retest and the tangential tracing retest.

[0090] After re-surveying 8 high-priority conflict zones, 21 out of 29 candidate boundary zones met the freezing conditions, forming 21 frozen boundaries; 5 were retained as unconfirmed boundaries due to partial occlusion and incomplete boundary bifurcation; and 3 were split into 7 boundary segments during the tracking process. Using the frozen boundaries as constraints, local topographic surface reconstruction was performed on the survey area, resulting in 14 independent reconstruction units. The total area of ​​the unconfirmed region is 612 m². 2 This portion, accounting for 6.4% of the total survey area, will not be included in the official survey results.

[0091] like Figure 5 As shown in the figure, this diagram selects a 30m control profile traversing B07 and compares the control profile, the unconstrained reconstructed profile, and the constrained reconstructed profile. In the figure, the green curve represents the control profile, the orange curve represents the unconstrained reconstruction result, and the blue curve represents the constrained reconstruction result. From... Figure 5 It can be seen that the unconstrained reconstruction exhibits significant smoothing near the boundary, resulting in weakened step edges; the constrained reconstruction, on the other hand, better preserves the boundary turning points and shows a significantly higher degree of fit with the control profile. Calculations show that the average absolute error of the unconstrained reconstruction relative to the control profile is 0.466m, and the root mean square error is 0.474m; the average absolute error of the constrained reconstruction is 0.066m, and the root mean square error is 0.068m, representing reductions of 85.8% and 85.7%, respectively.

[0092] The execution results of this embodiment show that the present invention does not uniformly remeasure all boundary objects. Instead, it first forms candidate boundary zones, and then determines the remeasurement priority based on cross-scale stability, the degree of separation between adjacent neighborhoods, and the impact of misjudgment. This allows remeasurement resources to be prioritized for the boundary areas that truly need confirmation. In this embodiment, the first round of remeasurement only targets eight high-priority conflict zones, achieving significant convergence of boundary deviations and enabling the final frozen boundary to directly constrain terrain surface reconstruction. The resulting mapping results not only maintain the true boundary morphology but also clearly distinguish unstable areas as areas to be confirmed, effectively reflecting the practical effects of the present invention in boundary identification and boundary constraint reconstruction.

[0093] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0094] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for optimizing the accuracy of geological regional surveying, characterized in that, The method includes: Acquire initial mapping data of the target geological area, and extract candidate boundary segments based on the initial mapping data; Candidate boundary bands are formed based on candidate boundary segments, and conflict zone priority sequences are generated based on the cross-scale stability index of the candidate boundary bands, the difference in feature distribution between the two neighboring areas, and the misjudgment impact value. The cross-scale stability index is used to indicate the stability of the candidate boundary bands at multiple scales, and the misjudgment impact value is used to indicate the degree of impact of misjudgment of the candidate boundary bands on region division and boundary reconstruction. Perform normal cross-traverse retesting and tangential tracing retesting according to the conflict zone priority sequence to determine the boundary position and update the candidate boundary zone; Based on the updated candidate boundary zones, constrained reconstruction is performed, and the mapping results of the target geological area are output.

2. The method according to claim 1, characterized in that, The step of extracting candidate boundary segments based on the initial mapping data includes: Based on the initial mapping data, determine the elevation change information, surface feature information, and observation integrity information; Candidate boundary segments are extracted based on the elevation change information, the surface feature information, and the observation integrity information.

3. The method according to claim 2, characterized in that, The step of forming candidate boundary bands based on candidate boundary segments includes: Based on the spatial proximity and directional continuity of the candidate boundary segments, the candidate boundary segments are spliced ​​together; If at least one of the elevation change information, the surface feature information, and the observation integrity information corresponding to adjacent candidate boundary segments satisfies the same trend, the splicing result is determined as a candidate boundary zone.

4. The method according to claim 1, characterized in that, The determination of the cross-scale stability index includes: Construct boundary saliency maps corresponding to the candidate boundary zones at multiple scales based on the initial mapping data; Extract the linear boundary skeleton based on each of the aforementioned boundary saliency maps; The cross-scale stability index is determined based on the continuous length, directional consistency, and morphological retention of the linear boundary skeleton at different scales.

5. The method according to claim 2, characterized in that, The determination of the differences in the distribution of features between the two neighboring areas includes: The first and second neighborhoods are determined along the normals of the candidate boundary bands, respectively. Geometric features in the first and second neighborhoods are determined based on the elevation change information; surface features in the first and second neighborhoods are determined based on the surface feature information; and observation integrity features in the first and second neighborhoods are determined based on the observation integrity information. The difference in feature distribution between the two neighboring regions is determined based on the statistical distribution differences between the first neighboring region and the second neighboring region on the geometric features, the surface features, and the observation integrity features; The geometric features include elevation residuals, slope, and roughness; the surface features include reflection intensity and texture features; and the observation integrity features include point density, occlusion rate, and porosity.

6. The method according to claim 1, characterized in that, The determination of the impact value of the misjudgment includes: The misjudgment impact value is determined based on the regional elevation difference, boundary length, area difference between adjacent regions, and boundary connection relationship corresponding to the candidate boundary zone.

7. The method according to claim 1, characterized in that, The generation of the conflict band priority sequence includes: Based on the cross-scale stability index, the difference in feature distribution between the two neighboring areas, the misjudgment impact value, and the observation sufficiency, the retest priority corresponding to the candidate boundary zone is determined. The candidate boundary bands are sorted according to the retest priority to generate the conflict band priority sequence; The observation sufficiency is used to characterize the degree of coverage of the candidate boundary zone by the current mapping data.

8. The method according to claim 1, characterized in that, Performing the normal crossing retest includes: For the target conflict zone in the conflict zone priority sequence, a crossing survey line is laid out along the normal direction of the target conflict zone; Obtain the remeasurement data corresponding to each of the aforementioned crossing survey lines; The boundary change points are determined based on the retest data, and the boundary positions are determined based on the differences in the distribution of local features on both sides of the boundary change points. The candidate boundary band is updated based on the boundary location.

9. The method according to claim 1, characterized in that, Performing the tangential tracing retest includes: For the target conflict zone in the conflict zone priority sequence, multiple overlapping tracking segments are deployed along the extension direction of the target conflict zone; Obtain the retest data corresponding to each of the aforementioned tracking segments; The continuous state, termination state, or segmented state of the candidate boundary band is determined based on the retest data corresponding to each tracking segment, and the candidate boundary band is updated based on the continuous state, termination state, or segmented state of the candidate boundary band.

10. The method according to claim 1, characterized in that, The constraint reconstruction based on the updated candidate boundary bands includes: The frozen boundary is determined based on the changes in the boundary position and the changes in the boundary connectivity of the updated candidate boundary zones; Using the frozen boundary as a boundary constraint, local terrain surface reconstruction is performed on the area enclosed by the frozen boundary; If the updated candidate boundary zone does not form a frozen boundary, the corresponding area will be identified as an area to be confirmed.

Images