Automatic production method and system for surveying and mapping section product and storage medium

Abstract

The invention provides an automatic production method and system for a surveying and mapping section product and a storage medium, and relates to the technical field of surveying and mapping engineering. The method comprises the following steps: loading DEM data generated by airborne laser radar point cloud and DOM data generated by an airborne camera; automatically laying section lines and checking directions; elevation points are extracted according to a set fixed or dynamically adjusted step length, two feature point extraction algorithms of a Douglas-Peucker algorithm and a Visvalingam-Whyatt algorithm are provided to adapt to different landforms and engineering scenes, and a minimum sampling interval is set; fusing elevation points actually measured by field workers by using RTK (Real Time Kinematic); the DOM and the DEM are superposed to generate a three-dimensional model, and interactive editing of DEM extraction section points and actually measured section points in a three-dimensional scene is supported; calculating mileage according to a water conservancy industry mileage calculation rule, and outputting standard water conservancy surveying and mapping section products in batches. According to the method, the problems of multiple redundant points, loss of feature points, low interaction efficiency and multi-software switching are solved, the section product precision is remarkably improved, and the working time is shortened by 80%.

Description

Technical Field

[0001] This invention relates to the field of surveying and mapping engineering technology, specifically to an automated production method, system, and computer-readable storage medium for surveying and mapping cross-section products based on lidar data, which is particularly suitable for cross-section surveying in fields such as water conservancy engineering and road surveying. Background Technology

[0002] In cross-section mapping operations in fields such as water conservancy projects and road surveying, traditional methods consistently face the dual challenges of efficiency and accuracy. The field survey phase requires manual collection of elevation data point-by-point along pre-defined cross-section lines using RTK equipment, which is not only time-consuming and labor-intensive but also prone to data blind spots due to terrain limitations. The data processing phase is equally cumbersome. Operators must plot the scattered points obtained in the field onto the CASS platform, manually draw the cross-section boundaries, select cross-section numbers and mileage extraction locations for each section, and then extract elevation points using a secondary development script and output them to a CSV file. This process not only involves an excessively high proportion of repetitive work but also introduces error risks due to the dense manual intervention, especially in complex terrain areas where the omission or mislabeling of feature points is frequent, resulting in significant deviations between the final cross-section map and the actual terrain.

[0003] With the widespread adoption of drone-based LiDAR technology, rapid acquisition of large-scale surface elevation data has become possible. However, significant shortcomings remain in the data processing stage: existing commercial software (such as AutoCAD plugins or professional LiDAR toolchains) has not yet formed an integrated solution, forcing users to repeatedly import and export data between multiple platforms—for example, first generating a DEM from point clouds and importing it into GIS software to define the scope, then manually laying out cross-sectional lines in the CAD environment, and finally relying on third-party tools to filter redundant points. This fragmented workflow not only reduces efficiency but also leads to a loss of accuracy due to data conversion. More importantly, the cross-section extraction process still relies heavily on manual operation, including steps such as forcibly drawing the range line, selecting cross-section numbers one by one, and extracting mileage values ​​one by one, with a serious lack of automation.

[0004] In addition, existing tools generally lack three-dimensional visualization capabilities. Operators can only make judgments based on elevation point symbols on two-dimensional planar views, making it difficult to intuitively identify terrain abrupt changes (such as steep slopes and gullies) or data anomalies in areas obscured by vegetation. Furthermore, the operation mode of switching between multiple software programs further exacerbates the learning cost and the rate of operational errors.

[0005] The aforementioned defects collectively lead to extended production cycles for surveying and mapping products and difficulty in ensuring the integrity of feature points, failing to meet the core requirements of modern engineering for "short construction period and high precision." Therefore, a new solution integrating automated processing and intelligent interaction is urgently needed. Summary of the Invention

[0006] I. Technical Issues 1. Eliminate manual intervention: Solve the problem of manually drawing range lines and selecting cross sections one by one when extracting cross section points; 2. Improve the efficiency of cross-sectional feature point extraction: Perform automated equidistant sampling and feature point extraction for cross-sectional lines; 3. Improve feature point accuracy: Enable real-time editing of cross-sectional points in 3D scenes; 4. Product standardization: Achieve automated output of standardized products.

[0007] II. Technical Solution To address the aforementioned technical problems, in a first aspect, the present invention provides an automated production method for surveying cross-section products, comprising the following steps: S1. Load the DEM generated from the point cloud of the airborne LiDAR, the DOM generated from the airborne camera, and the section line design file; S2. Cross-section layout and inspection: Automatically check the direction of cross-section lines based on a two-dimensional planar map, support custom cross-section names, and automatically number the cross-section lines; S3. Batch extraction and filtering of cross-section points: Based on DEM data, DEM elevation points are extracted along the cross-section line at a preset step size, and feature points are extracted using two algorithms: Douglas-Peucker and Visvalingam-Whyatt. S4. Data Fusion: Import RTK measured elevation feature point data and fuse it with automatically extracted feature points; S5. 3D Interactive Modification: Overlays DOM and DEM to generate 3D terrain, automatically extracts and measures elevation feature points from the overlay, edits cross-section points in the 3D scene, and supports automatic annotation of feature points; S6. Batch output of results: Automatically calculate cross-section mileage and export cross-section diagrams in DWG format and CSV / DAT data tables.

[0008] Preferably, the DEM generation in step S1 includes: Denoising and classification of lidar point clouds; High-precision DEM is generated by interpolation of irregular triangular mesh (TIN).

[0009] Preferably, when extracting DEM elevation points in step S3, the sampling step size can be set by the user to a fixed step size, or set to a step size that is dynamically adjusted according to the average density of the lidar point cloud. When the step size is set to be dynamically adjusted, the rules include: Density > 10 points / m 2 The time step length is ≤1.0m; Density ≤ 10 points / m 2 The time step length is ≤2.0m.

[0010] Preferably, the filtering algorithm in step S3 includes: Douglas-Peucker sets the default tolerance parameter to 0.3m, while Visvalingam-Whyatt sets the default tolerance parameter to 0.1m. 2 Preserve terrain feature points whose curvature changes exceed a threshold; Preferably, step S4 includes: Based on the real terrain and landform features reflected in the 3D scene, the location of feature points is determined, and the elevation feature points are manually corrected and supplemented. It also supports the import and automatic annotation of measured feature points.

[0011] Preferably, the cross-sectional mileage rule in step S5 is as follows: Facing downstream, the left side represents the starting point of the negative mileage, and the right side represents the ending point of the positive mileage.

[0012] In a second aspect, the present invention provides an automated production system for surveying cross-section products, applicable to the method described in the first aspect above, comprising: The data input unit is used to input the DEM generated by the airborne LiDAR point cloud, the DOM generated by the airborne camera, and the cross-section design file; The cross-section layout unit is connected to the data input unit and configured to lay out cross-section lines on a two-dimensional planar map and automatically number them. The feature extraction unit and the cross-section layout unit are configured as follows: ① Extract elevation points along the cross-section line according to the preset step size; ② The Douglas-Peucker and Visvalingam-Whyatt algorithms were used to extract elevation feature points; The 3D interactive unit connects to the feature extraction unit and is configured to overlay the DOM and DEM to generate 3D terrain, supporting cross-section point editing. The output unit connects to the 3D interactive unit and is configured to automatically calculate mileage and export results in batches.

[0013] Preferably, the three-dimensional interactive unit includes: The display and editing sub-unit is configured to overlay automatically extracted elevation feature points and measured points on the real 3D terrain generated using DEM and DOM, and respond to user operations to add or delete cross-section points and label feature points.

[0014] Thirdly, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0015] III. Compared with the prior art, the beneficial effects of the present invention are: This invention systematically solves the core pain points of traditional cross-section surveying by constructing a full-link technical framework of "data-driven, intelligent extraction, and 3D interaction." At the data input level, the DEM (Digital Elevation Model) data generated by airborne LiDAR point clouds and the DOM (Digital Orthophoto) data generated by airborne cameras eliminate multiple software conversion steps, improving data consistency from the source. At the feature point extraction level, by providing two flexible sampling modes—fixed step size and dynamically adjustable step size—users are empowered to balance efficiency and accuracy according to actual needs. Combined with the subsequent high-efficiency Douglas-Peucker and Visvalingam-Whyatt intelligent filtering algorithms, the feature point capture rate is successfully increased to over 95%. In the interaction stage, innovative integration, along with real-time editing tools, improves cross-section correction efficiency by 50%. Simultaneously, the downstream-oriented left-negative, right-positive mileage rule and batch output engine work together to reduce the surveying cycle of a 10-kilometer river cross-section from the traditional 3 days to 4 hours.

[0016] Ultimately, this achieves a triple breakthrough in accuracy, efficiency, and user experience: elevation errors are stably controlled within specified limits; manual operation steps are reduced by 80%; and the feature point omission rate in complex terrain is reduced to below 5%. It truly realizes an industrialized surveying and mapping production paradigm of "one-click input, fully automated processing, 3D visualization quality inspection, and standardized output." Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0018] Figure 1 This is a schematic diagram of the user interface used for loading raw data such as DEM, DOM, and cross-section lines in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the direction check of imported cross-section lines on a two-dimensional planar map in an embodiment of the present invention; Figure 3 This is a schematic diagram of the user interface used to set parameters for batch extraction, filtering, and fusion of measured data of cross-section points in an embodiment of the present invention. Figure 4 This is a schematic diagram of the interactive modification main interface for the integrated linkage of "figure-table-3D" in an embodiment of the present invention; Figure 5 for Figure 4 The interface shown is a detailed illustration of highlighting and locating a single feature point for fine-tuning. Figure 6 This is a schematic diagram of a user interface for setting batch output parameters of results in an embodiment of the present invention; Figure 7This is a schematic diagram showing a list of various format result files generated in batches after the method of this embodiment of the invention is executed. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] Example 1: This embodiment provides an automated production method for cross-section mapping products, aiming to systematically address a series of long-standing pain points in existing technologies, such as cumbersome workflows, low automation, difficulty in guaranteeing feature point extraction accuracy, low human-computer interaction efficiency, and insufficient standardization of output results. This embodiment significantly improves the production efficiency and quality of cross-section mapping by constructing a complete, integrated software processing flow from raw data input to final standardized output.

[0021] The first step, S1, involves loading the Digital Elevation Model (DEM) generated from the airborne LiDAR point cloud, the Digital Orthophoto Map (DOM) generated from the airborne camera, and the pre-planned cross-section design file. This initial step forms the data foundation for the entire automated process, and its execution quality directly affects the accuracy and efficiency of all subsequent steps.

[0022] Specifically, operators can use methods such as Figure 1The system data import interface shown allows users to select and load all necessary raw data at once. During execution, the system optionally first performs a series of refined preprocessing operations on the raw airborne LiDAR point cloud, which serves as the core data source. These preprocessing operations include, but are not limited to, denoising the LiDAR point cloud. For example, the system can use statistical outlier removal algorithms to identify and remove obvious noise points and drift points generated during flight or scanning, or use isolated point denoising algorithms to remove scattered, unsupported isolated points in the point cloud, thus obtaining a cleaner and purer raw point cloud dataset. Next, the system performs precise classification processing on the denoised point cloud data. This step is crucial; the system can employ advanced ground point filtering algorithms, such as a progressively encrypted triangular mesh filtering algorithm, to accurately classify the massive point cloud data into different categories such as ground points, vegetation points, building points, and water body points. In the subsequent terrain model generation, to ensure that the DEM accurately reflects the surface morphology, the system will primarily process the precisely classified ground point data. After classifying the point cloud, the system utilizes Triangular Irregular Network (TIN) interpolation technology to construct and generate a high-precision Digital Elevation Model (DEM) covering the entire survey area based on these rigorously selected ground points. This DEM model is the fundamental basis for all subsequent elevation information extraction. To adapt to the diverse accuracy requirements of different engineering projects, the grid spacing of the DEM generated in this embodiment is selectable by the user. For example, the user can choose to generate DEM products with different resolutions such as 0.1 meters, 0.3 meters, or 0.5 meters according to actual needs.

[0023] While loading the DEM, to provide realistic and intuitive geographic background information to assist subsequent human-computer interaction, the system will simultaneously load one or more digital orthophoto maps (DOMs) that are spatially fully registered with the DEM data. Loading the DOM allows operators to intuitively refer to the real textures and colors of landforms to determine terrain features in subsequent interactive processes.

[0024] In addition, the system supports loading cross-section design files containing preset cross-section layout information. To maximize compatibility with existing design workflows and data standards, the system supports multiple mainstream vector data formats. For example, users can directly import AutoCAD's DWG or DXF format files, or the SHP format files commonly used in the GIS field. All loaded data, whether raster format DEMs and DOMs or vector format cross-section files, will be precisely geoaligned and spatially correlated within the system using a unified, high-precision coordinate system. This data alignment strategy from the source ensures that no cumbersome coordinate system conversions or data format transfers are required at any step in the subsequent processing flow, thus fundamentally guaranteeing the inherent consistency, high efficiency, and high accuracy of the data processing workflow.

[0025] After the basic data preparation is completed, the cross-sections are laid out and checked. This step is crucial to ensuring that the cross-section layout meets surveying specifications and design requirements, and all operations are performed within the two-dimensional planar map window provided by the system.

[0026] like Figure 2 As shown, the window displays the loaded DOM as a clear base map, overlaying the imported cross-sectional lines as vectors, providing operators with an intuitive inspection environment. The system automatically and intelligently performs a rigorous directional check on each imported cross-sectional line. Especially in application scenarios such as water conservancy projects where the orthogonality of cross-sections is strictly required, the system automatically calculates the actual angle between each cross-sectional line and the main orientation line, such as the pre-imported or identified river centerline or road centerline. If the system detects that the angle of a cross-sectional line deviates from the ideal 90-degree orthogonality, and the deviation exceeds a threshold that can be preset by the user (e.g., 5 degrees), the system will immediately trigger an early warning mechanism. At this time, a prominent red highlighted warning bar will pop up on the interface, providing a clear suggested rotation angle, thus providing operators with clear and timely correction instructions. Operators can choose "one-click confirmation" to have the system automatically complete the rotation correction based on the system prompts, or they can choose to enter the manual fine-tuning mode to make more precise adjustments to the angle of the cross-section line until it fully meets the requirements of the design specifications.

[0027] After ensuring the accuracy of the direction, the system will further automatically determine and mark the left and right directions of each cross-section line according to the mileage calculation rules widely used in the water conservancy industry, which always face the downstream direction of the river. The correctness of this directionality directly affects the accuracy of the subsequent mileage calculation results. For visualization, this direction will be marked directly on the cross-section line of the two-dimensional map with a clear arrow symbol. If the operator finds that the direction marking of a cross-section line is incorrect during the inspection process, they can quickly and conveniently reverse its direction by simply double-clicking the cross-section line with the mouse.

[0028] Once all orientation checks and corrections are complete, the system will initiate an automated batch numbering function, assigning unique and consecutive numbers to all cross-section lines. To meet the management needs of different projects and stages, the system offers a highly flexible custom numbering rule function. For example, operators can freely set the character prefix for the numbers according to the specific requirements of the project, such as using "DM" to represent "section" or "CS" to represent "cross section"; they can also set the starting number, such as starting sequentially from "1" or from any specified number like "100". This ensures that the cross-section numbers in the final system-generated office documents are completely consistent with the record numbers used by the project team during field surveys, thus achieving seamless integration of office and field data management.

[0029] Next, this embodiment will proceed to step S3, which involves batch extraction and intelligent filtering of cross-sectional points. For example... Figure 3 As shown, its detailed parameter setting interface provides operators with a high degree of flexibility. This step represents a key technological breakthrough in realizing the transition from manual to automated cross-section production.

[0030] The system will automatically and in batches extract the elevation point information recorded in the DEM model below each cross-section line that has been rigorously checked and assigned a unique number in step S2. To adapt to the accuracy and efficiency requirements of different engineering projects, this system provides users with two flexible step size setting modes in the DEM sampling stage: fixed step size mode and dynamic step size mode.

[0031] 1. Fixed step size mode: As attached Figure 3 As shown, operators can directly set a specific, fixed value (e.g., 3.0 meters) in the "Sampling Step Size" input box on the "Cross-Section Extraction" interface. In this mode, the system will strictly follow the set value to sample elevation points at equal intervals along each cross-section line. This mode is simple and intuitive, ensuring a uniform distribution of sampling points, and is suitable for application scenarios where there are uniform requirements for the spacing of data points.

[0032] 2. Dynamic Step Size Mode: As a more intelligent option, users can also choose to enable dynamic step size mode. This mode adaptively adjusts the sampling step size of elevation points in real time based on the average density of LiDAR point cloud data in local areas along the cross-section. Specifically, when the system detects that the point cloud density in the area traversed by the cross-section is very high, such as greater than 10 points / square meter, which usually indicates that the terrain in that area is relatively complex or critical, the system will automatically compress the sampling step size to a smaller value, such as ≤1.0 meter, to capture more terrain details. Conversely, when the point cloud density in the area traversed by the cross-section is low, such as ≤10 points / square meter, which usually corresponds to a relatively flat and open terrain, the system will automatically widen the sampling step size to a relatively larger value, such as ≤2.0 meters.

[0033] By offering these two modes, users are given full choice, enabling them to achieve the best balance between processing efficiency and result accuracy according to their actual needs.

[0034] After extracting an initial, relatively dense sequence of elevation points using a dynamic step-size strategy, the system immediately activates a subsequent intelligent filtering algorithm to finely screen these initial point sequences, as this equidistant or near-equidistant sampling method inevitably introduces a large number of redundant data points that are not very meaningful for describing terrain transitions. The goal is to accurately extract the most representative terrain feature points that can most realistically reflect the transitions in landforms from a large number of sampled points.

[0035] In order to flexibly adapt to the complex surveying needs of different landform features, this embodiment provides users with at least two advanced filtering algorithms that are widely recognized in the industry and have different characteristics for selection.

[0036] The first algorithm is the classic Douglas-Peucker algorithm. Users can set a default tolerance parameter for this algorithm, for example, setting it to 0.3 meters. The core advantage of the Douglas-Peucker algorithm lies in its ability to excellently preserve feature points with significant curvature changes in the terrain. Therefore, it is particularly suitable for preserving cross-sectional features of artificial features with obvious geometrically sharp inflections, such as the upper and lower edges of steep slopes, the crest lines of embankments, the edges of artificial steps, and the edges of buildings.

[0037] The second algorithm is the Visvalingam-Whyatt algorithm. Similarly, users can set a default tolerance parameter for this algorithm, for example, 0.1 square meters. Unlike the Douglas-Puk algorithm, the Visvalingam-Whyatt algorithm, when filtering points, tends to maintain the overall, gentle curvature of the terrain, effectively avoiding unrealistic landform distortions caused by oversimplification in some gentle terrain transition areas. Therefore, this algorithm is very suitable for processing natural landform sections with smooth transitions, such as natural riverbanks and contour lines of mountains.

[0038] By providing users with two complementary algorithm options, operators can directly choose based on their professional experience, taking into account the actual working conditions and terrain type; alternatively, they can conduct experimental comparisons by running the two algorithms separately and ultimately retain the result that produces the better extraction effect. This flexible mechanism ensures that the final extracted terrain feature points can accurately retain all key terrain inflection points with curvature changes exceeding the user's preset threshold, while also conforming to the actual terrain morphology to the greatest extent possible.

[0039] After completing automated feature point extraction and intelligent filtering, to further ensure the absolute accuracy of the final results at key locations and effectively compensate for the potential limitations of relying solely on LiDAR data, this embodiment specifically introduces step S4, namely, fusing field measurement data. Considering that LiDAR technology may inherently suffer from insufficient accuracy or complete data loss in certain specific and challenging areas, such as underwater topography measurement, precise positioning at the edges of structures, or obtaining surface elevation in areas with extremely dense vegetation, this embodiment designs a function to fuse high-precision field measured elevation points. This step greatly improves the overall quality and reliability of the final results.

[0040] In specific implementation, such as Figure 3As shown in the "Include Measured Points" checkbox, operators can activate this function and then specify the path to an external data file containing the 3D coordinate information of the measured points in the "Measured Point Data Path" field. The format is typically an industry-standard .dat or .csv file. After data import, the system will prompt the user to set a crucial parameter: the "Measured Point Fusion Distance Threshold." For example, the user can flexibly set this threshold to 1.0 meter based on the actual measurement accuracy and terrain complexity. Once the parameter is set, the system will initiate an automated fusion processing procedure. This procedure efficiently iterates through each imported measured point and accurately calculates the shortest (usually perpendicular) spatial distance between that point and all cross-sectional lines within the engineering scope. If the system calculates that the distance between a measured point and a specific cross-sectional line is less than the user-preset fusion distance threshold, then that measured point will be automatically identified by the system as a valid, high-precision supplementary point for that cross-section.

[0041] Subsequently, the system performs a fusion operation: it retains the original high-precision measured elevation value of the measured point provided by equipment such as RTK, and recalculates the accurate cross-sectional mileage based on the point's precise projection position on the target cross-section line. In this way, the high-precision measured point is seamlessly and perfectly integrated into the feature point sequence of the cross-section while maintaining its elevation accuracy.

[0042] This fusion mechanism ingeniously combines the advantages of lidar technology—its wide coverage, high efficiency, and high density—with the unparalleled high precision of traditional field surveying methods such as RTK at key points. Ultimately, it ensures that the final cross-section results have absolute accuracy at all critical terrain feature points essential to engineering design, thereby fundamentally improving the overall quality and engineering practical value of the surveying cross-section products.

[0043] After all automated data processing and high-precision measured data fusion are completed, this embodiment enters a crucial, manually-led quality control stage, namely step S5, which involves interactive 3D visualization modification. This is a key guarantee that the automated processing results fully meet the design intent and that any minor imperfections are finely corrected. To provide operators with the most intuitive and realistic inspection and editing environment, the system first performs a 3D scene construction task. It precisely and seamlessly overlays the DOM (Digital Orthophoto) loaded in step S1, as a high-fidelity texture, onto the DEM (Digital Elevation Model) also generated in step S1. Through this texture mapping technology, the system can dynamically generate a 3D terrain scene with realistic landform textures, colors, and spatial accuracy within a 3D scene.

[0044] Its core integrated interactive interface, combining graphs, tables, and 3D models, is as follows: Figure 4 As shown. In this highly realistic 3D scene, the system further overlays and visualizes all the key data points obtained in the previous steps. This includes the elevation feature points automatically extracted by the intelligent algorithm in step S3, which are rendered with a specific color, such as blue, for easy differentiation. Meanwhile, the high-precision measured points added in step S4 by fusing field data can be rendered with another striking color, such as green, to clearly indicate their data source and high-precision attributes.

[0045] Furthermore, by combining the DEM with the DOM to generate a high-fidelity 3D terrain model for visualization, there is no need to directly load and display the original, massive LiDAR point cloud. This not only significantly reduces the consumption of computer hardware performance and ensures the smoothness of 3D scene interaction, but also avoids visual clutter caused by excessively dense point cloud data. This allows operators to focus entirely on key terrain features and cross-sectional points, enabling more efficient and accurate editing and quality checks. This multi-source data fusion display allows operators to comprehensively judge the macroscopic trends and microscopic details of the terrain in a unified view. Users can freely rotate, scale, and smoothly navigate the 3D scene from any angle using very simple and intuitive mouse drag-and-drop operations, enabling them to quickly and accurately locate any cross-section that needs to be inspected.

[0046] When manual intervention and fine-tuning are required, operators can directly add missing terrain feature points at any location on the cross-section line in this 3D window by clicking with the mouse, or remove redundant or unreasonable feature points that may have been generated during automated processing by selecting and deleting them. Furthermore, users can even fine-tune the 3D spatial position of a feature point by dragging with the mouse to make it more perfectly match the actual terrain transitions.

[0047] The core interaction mechanism of this embodiment is its integrated "graph-table-3D" linkage function, such as... Figure 4 As shown. Modifications made by the operator in any view of the system interface, whether in a 3D scene or... Figure 4 The two-dimensional cross-sectional view shown below, as well as the cross-sectional list on the left, will be automatically and synchronously updated to the other two views in real time without delay. For example, when the user is in... Figure 4 When you click on a row in the "Cross-section List" on the left, as shown Figure 5As shown, the corresponding cross-section line in the 3D scene will be immediately highlighted, and the view will automatically focus on the cross-section line, greatly facilitating users to quickly locate and inspect it. Similarly, users can add detailed attribute information to a feature point in the 3D scene via the right-click menu. For example, it can be labeled as "embankment shoulder," "waterside line," or "embankment." This valuable attribute labeling information will be immediately updated and associated with the corresponding entry in the cross-section list, and will eventually be completely output to the final result file, with automatic labeling on the corresponding feature point in the CAD cross-section drawing.

[0048] This WYSIWYG, highly integrated editing approach significantly improves the efficiency and accuracy of the final correction of cross-sectional data. To ensure data security and consistency, all manual editing and modification operations, including adding, deleting, and modifying points, as well as labeling attributes, are written back to the backend in real time and synchronously for persistent storage in a professional geospatial database such as PostgreSQL. This mechanism ensures high data consistency throughout the entire processing flow and effectively guarantees that no further manual verification or data synchronization is required when outputting the final results, thus achieving a truly seamless workflow.

[0049] After all the cross-sectional data have undergone the aforementioned rigorous and meticulous checks, and have been finally confirmed by the operators to be error-free, this embodiment proceeds to the final stage, step S6, which involves the batch output of standardized results. This is a crucial step in transforming all intermediate processing results into the final deliverable. In this step, the system first strictly adheres to the standardized mileage calculation rules recognized within the water conservancy industry, namely, "when facing downstream of the river, the mileage value of all points on the left side of the cross-section line is negative, the mileage value of all points on the right side is positive, and the absolute value of the mileage increases from the central base point towards both banks." For each cross-section, the system automatically and accurately calculates the mileage of all elevation feature points on the cross-section line. The mileage calculation accuracy can reach 0.01 meters.

[0050] After the mileage calculation is fully automated, the operator only needs to... Figure 6In the batch output interface shown, by selecting the required file types and setting necessary parameters such as horizontal and vertical scales, project name, and elevation system, and then clicking the "Export" button, the system immediately launches a concurrent processing engine based on multi-threading technology. This engine can intelligently distribute a large number of cross-section processing tasks across multiple processor cores for efficient and parallel processing. This means that the system can process multiple cross-sections simultaneously, thereby greatly reducing the waiting time for batch output. Ultimately, the system can generate a complete, highly standardized, and directly usable set of water conservancy engineering surveying cross-section products in one click and in batches.

[0051] like Figure 7 As shown, the output folder clearly lists the generated files in various formats, such as file20250723_cross-section plot.dxf, file20250723_cross-section.csv, and file20250723_cross-section feature points.dat. To meet the complex needs of different stages and downstream applications in engineering projects, the system outputs results in multiple mainstream data formats.

[0052] First and foremost, one of the most crucial achievements is that the system generates cross-sectional and longitudinal profile views in DWG format that conform to industry drafting standards. All elements of these drawings, including the style and size of the drawing frame, the drawing scale (the system provides a rich selection of scales, including but not limited to commonly used scales such as 1:200, 1:500, and 1:1000), the style and size of various annotation texts, and the layer division of all graphic elements, are strictly preset and automatically generated according to the design and drafting specifications of the water conservancy and hydropower industry. This means that these DWG files can be used directly in the design, review, and final drawing stages without any secondary editing.

[0053] Secondly, to facilitate further data analysis and utilization, the system simultaneously generates data table files in CSV and DAT formats. CSV files are general-purpose comma-separated text files that can be easily imported into spreadsheet software such as Microsoft Excel for various complex statistical analyses, data visualizations, or secondary processing. DAT files, on the other hand, are a more professional data format. The DAT files generated by the system strictly adhere to the cross-sectional data format standards of software that dominates the Chinese surveying and mapping industry, such as CASS. This means that the width and order of its fields are fixed and standardized, allowing for seamless and direct use by subsequent professional software such as earthwork volume calculation, 3D terrain modeling, and hydraulic model analysis, avoiding errors and inconveniences that may arise from data format conversion.

[0054] Furthermore, to enhance the flexibility of output and adapt to the differentiated needs of output at different project stages, this embodiment also features a highly flexible batch output option. For example, operators do not need to output all cross-sections at once. Instead, they can choose to output outputs by cross-section number range (e.g., only cross-sections from DM100 to DM200), or by the geographical segmentation of the river channel (e.g., only cross-sections within a specific section), or even by using custom, more complex filtering conditions (e.g., only outputting cross-sections with specific attribute annotations). This high degree of flexibility allows this embodiment to perfectly adapt to the diverse needs for cross-section outputs throughout different project cycles, from the preliminary design stage to the detailed construction drawing design stage, and then to the later operation and management stage.

[0055] In summary, the automated production method for surveying cross-section products provided in this embodiment systematically constructs a complete and closed-loop automated production process, from loading and preprocessing raw data, to setting up and inspecting cross-sections, to intelligent extraction and filtering of feature points using multiple algorithms and fusion of measured data, to immersive 3D visualization and interactive modification, and finally to the batch production and one-click output of standardized results. This successfully and comprehensively solves many long-standing pain points in the traditional surveying cross-section production operation mode.

[0056] When using this method, the workflow of surveying engineers is greatly simplified and optimized: the operator only needs to load all the raw data into the system corresponding to this method in one convenient way; then, in the powerful interactive environment provided by the system that integrates "two-dimensional cross-sectional views and three-dimensional real scenes", the results generated by the system's automated processing are efficiently, intuitively, and accurately checked for quality and necessary fine manual modifications; finally, with just one click, all final results that meet industry design and delivery specifications can be output in batches.

[0057] This method has been thoroughly applied and rigorously verified in multiple practical engineering projects, including a large-scale irrigation area renovation project in the Huai River Basin. In terms of production efficiency, compared to the traditional work mode that relies on multiple software programs and extensive manual operations, this method reduces manual operation steps by more than 80%. The complete surveying cycle for a typical 10-kilometer-long river cross-section, which previously required at least 3 working days, has been reduced to less than 4 hours, representing an efficiency improvement of nearly an order of magnitude. In terms of accuracy and reliability, by introducing a collaborative working mechanism of multiple intelligent filtering algorithms and seamlessly integrating with high-precision RTK field measurement data, this method successfully reduced the omission rate of key feature points in complex terrain (such as steep slopes and valleys) from more than 20% in traditional methods to less than 5%, and the feature point capture rate was correspondingly increased to more than 95%. The elevation error of the final result can be stably controlled within the standard limit, which makes the first-time acceptance rate of the result jump from about 78% to more than 98%. Furthermore, the consistency between the generated cross-section shape and the cross-section shape measured in the field is as high as 95%, which fully guarantees the authenticity of the result.

[0058] Ultimately, through its integrated and intelligent design of the entire process, this method truly achieves the ideal goal of "one-click input, fully automatic processing, three-dimensional visualization quality inspection, and standardized output" in the field of surveying and mapping section production. This forms a new, efficient, accurate, and modern industrialized surveying and mapping production paradigm, providing strong support for the technological progress of the entire industry.

[0059] Example 2: This embodiment provides an automated production system for surveying cross-section products. This system serves as the physical and logical carrier of the automated production method for surveying cross-section products detailed in Embodiment 1. This system is applicable to executing the method described in Embodiment 1. Its core feature is that it structurally includes a data input unit, a cross-section layout unit, a feature extraction unit, a 3D interaction unit, and an output unit. These five core functional units are logically tightly coupled and work collaboratively to ensure the smooth and efficient execution of the entire automated processing flow. In terms of physical deployment, this system offers excellent flexibility and applicability. Its core design aims to meet the needs of efficient operation in a standalone environment, seamlessly integrating and deploying all functional units on a single computing device, especially running smoothly on mainstream laptops, thus forming a fully functional and highly portable standalone system. This feature greatly facilitates the flexible use of surveying personnel at project sites or in offices. Simultaneously, for large-scale collaborative operations or online processing scenarios with special requirements, this system can also be deployed in a distributed manner using a client-server architecture. Regardless of the deployment method, these units exchange data and communicate commands in real time through internal buses or networks, thereby ensuring seamless connection and efficient operation of the entire technical process.

[0060] Specifically, the data input unit is the starting point of the entire system. Its main function is to input the DEM generated by the airborne LiDAR point cloud, the DOM generated by the airborne camera, and the cross-section design file. Its user interface is as described above. Figure 1 As shown. In terms of physical implementation, this unit is typically deployed within a high-performance graphics workstation host, equipped with a high-speed NVMe solid-state drive array to ensure rapid reading of large-capacity point cloud and image data, such as LAS, LAZ, or XYZ format point cloud files larger than 500GB per file. This unit incorporates parallel generation submodules for DEM and DOM, enabling collaborative computation using multi-core CPUs and GPUs. For example, by parallelizing TIN interpolation and rasterization processes, its processing speed can be more than ten times faster than traditional single-threaded methods, effectively avoiding disk I / O bottlenecks in the data loading and preprocessing stages.

[0061] The cross-section layout unit is logically connected to the data input unit to obtain the loaded DOM and cross-section line data. This unit is configured to lay out cross-section lines on a two-dimensional planar map and automatically number them; its working scenario is as follows: Figure 2 As shown. In terms of physical implementation, the user interface of this unit can run on a standalone terminal device, such as a 24-inch dual-touchscreen terminal, communicating with the host via Gigabit Ethernet or 802.11ac Wi-Fi, facilitating temporary deployment and collaborative work in the field. Operators can conveniently draw cross-section lines, check directions, and set numbering rules on the terminal's touchscreen.

[0062] The feature extraction unit is connected to the cross-section layout unit to receive the cross-section line information after layout and inspection. This unit is configured to: first, extract elevation points along the cross-section line at dynamic step sizes; second, extract elevation feature points using algorithms such as Douglas-Peucker and Visvalingam-Whyatt, the relevant parameter setting interface of which is [insert interface here]. Figure 3 In terms of physical implementation, this unit can share the graphics processing unit (GPU) configured on the device with the 3D interactive unit, leveraging its parallel computing capabilities to accelerate the extraction of elevation feature points. On laptops or workstations equipped with dedicated graphics cards, this allows rendering calculations and feature extraction algorithms to be performed efficiently without requiring extensive data copying.

[0063] This unit integrates a step size setting module and a filtering algorithm selection module: the step size setting module is used to implement the user-selected fixed step size sampling, or to enable the dynamic step size adjustment function based on point cloud density; the filtering algorithm selection module supports switching between different feature point extraction algorithms such as Douglas-Peucker and Visvalingam-Whyatt.

[0064] The 3D interactive unit connects to the feature extraction unit to receive the extracted feature point data. This unit is configured to overlay the DOM and DEM to generate 3D terrain and supports editing of cross-sectional points within the 3D scene. Its complete interactive interface is shown below. Figure 4 As shown, this unit includes a 3D scene, cross-sectional views, and an attribute list. The physical implementation of this unit typically includes a high-performance graphics workstation, a professional graphics card, and a high-resolution monitor, such as a 32-inch 4K stereoscopic display, providing users with an immersive, six-DOF stereoscopic visual roaming and editing experience. Specifically, this unit includes a display and editing sub-unit, which clearly overlays all automatically extracted elevation feature points and fused measured points onto a 3D terrain with realistic textures generated by overlaying the DEM and DOM, providing users with a comprehensive, WYSIWYG visualization and interactive environment. This unit responds to user operations, such as adding, deleting, and repositioning cross-sectional points, and annotating attributes through mouse or 3D mouse clicks and drags. Figure 5 The image shows the precise location and highlighting of a single point.

[0065] Finally, the output unit connects to the 3D interactive unit to obtain all cross-sectional data that has been finalized by the user. This unit is configured to automatically calculate mileage and batch export the final results, offering a wide range of export options such as... Figure 6As shown. In terms of physical implementation, this unit can be deployed on a backend server node, integrating an AutoCAD core engine to ensure the highest compatibility of the generated DWG files. The unit connects to a high-capacity RAID6 storage array via a high-speed network, such as a 10Gbps fiber optic channel, ensuring sufficient storage capacity to meet the online storage needs of all project results within a year. When a user triggers a batch export operation, the unit first calculates the total mileage according to preset mileage rules, then starts a multi-threaded concurrent engine to process each section independently, ultimately generating batches of files as shown. Figure 7 The output files are shown in various formats.

[0066] The system provided in this embodiment, through the clear division of labor and close collaboration among its functional units, solidifies the methods described in Embodiment 1 into a complete and efficient hardware and software solution. From the high-speed data loading and preprocessing of the data input unit, to the convenient human-computer interaction of the cross-section layout unit, to the intelligent algorithm processing of the feature extraction unit, and the immersive and intuitive editing and quality inspection environment provided by the 3D interaction unit, and finally to the batch and efficient generation of standardized results by the output unit, the design of the entire system ensures automation and integration of the entire process from raw data to final product.

[0067] The system is flexible in its deployment, serving as both a standalone high-performance workstation and a collaborative work platform under a client / server architecture. Its hardware and software configuration fully considers the performance requirements of large-scale data processing, providing solid technical support for the efficient, accurate, and industrialized production of surveying and mapping profile products.

[0068] Example 3: This embodiment also provides a computer-readable storage medium on which computer program code is permanently stored. Its core feature is that, when the computer program is loaded and executed by one or more processors, it can completely and accurately implement all the steps of the detailed automated production method for surveying cross-section products described in Embodiment 1, and can generate a user interface and final results as shown in the accompanying drawings.

[0069] The computer-readable storage medium can be any physical device capable of storing program code, such as, but not limited to, a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a solid-state drive (SSD), a magnetic disk, or an optical disk.

[0070] When the device loaded with the computer program is running, the processor reads and executes the instructions stored on the medium, thereby controlling the device to complete a complete and automated processing flow. This flow covers all the steps from data loading and preprocessing, to interactive layout and automated inspection of cross sections, to feature point extraction and filtering based on multiple intelligent algorithms and the fusion of measured data, immersive 3D visualization and interactive modification, and finally to the batch production and one-click output of a series of standardized results.

[0071] In this way, the computer program stored on the medium enables any general-purpose computing device that meets the hardware requirements to be transformed into a powerful tool for automated production of surveying cross sections, thereby comprehensively and efficiently achieving all the expected technical effects of this invention.

[0072] Although the present invention has been described in detail above with general descriptions and specific embodiments, some modifications or improvements can be made to it, such as reasonable changes to parameters like point cloud density, tolerance parameters, and output format, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. An automated production method for surveying cross-section products, characterized in that, Includes the following steps: S1. Load the DEM generated from the point cloud of the airborne LiDAR, the DOM generated from the airborne camera, and the section line design file; S2. Cross-section layout and inspection: Automatically check the direction of cross-section lines based on a two-dimensional planar map, support custom cross-section names, and automatically number the cross-section lines; S3. Batch extraction and filtering of cross-section points: Based on DEM data, DEM elevation points are extracted along the cross-section line at a preset step size, and feature points are extracted using two algorithms: Douglas-Peucker and Visvalingam-Whyatt. S4. Data Fusion: Import RTK measured elevation feature point data and fuse it with automatically extracted feature points; S5. 3D Interactive Modification: Overlays DOM and DEM to generate 3D terrain, automatically extracts and measures elevation feature points from the overlay, edits cross-section points in the 3D scene, and supports automatic annotation of feature points; S6. Batch output of results: Automatically calculate cross-section mileage and export cross-section diagrams in DWG format and CSV / DAT data tables.

2. The automated production method for surveying cross-section products according to claim 1, characterized in that, The DEM generation in step S1 includes: Denoising and classification of lidar point clouds; High-precision DEM is generated by interpolation of irregular triangular mesh (TIN).

3. The automated production method for surveying cross-section products according to claim 1, characterized in that, When extracting DEM elevation points in step S3, the sampling step size can be set by the user to a fixed step size, or set to a step size that is dynamically adjusted according to the average density of the lidar point cloud. When the step size is set to be dynamically adjusted, the rules include: Density > 10 points / m 2 The time step length is ≤1.0m; Density ≤ 10 points / m 2 The time step length is ≤2.0m.

4. The automated production method for surveying cross-section products according to claim 1, characterized in that, The filtering algorithm in step S3 includes: Douglas-Peucker sets the default tolerance parameter to 0.3m, while Visvalingam-Whyatt sets the default tolerance parameter to 0.1m. 2 Topographic feature points with curvature changes greater than a threshold are retained.

5. The automated production method for surveying cross-section products according to claim 1, characterized in that, Step S4 includes: Based on the real terrain and landform features reflected in the 3D scene, the location of feature points is determined, and the elevation feature points are manually corrected and supplemented. It also supports the import and automatic annotation of measured feature points.

6. The automated production method for surveying cross-section products according to claim 1, characterized in that, The cross-sectional mileage rule in step S5 is as follows: Facing downstream, the left side represents the starting point of the negative mileage, and the right side represents the ending point of the positive mileage.

7. An automated production system for surveying cross-section products, applicable to the method described in any one of claims 1-6, characterized in that, include: The data input unit is used to input the DEM generated by the airborne LiDAR point cloud, the DOM generated by the airborne camera, and the cross-section design file; The cross-section layout unit is connected to the data input unit and configured to lay out cross-section lines on a two-dimensional planar map and automatically number them. The feature extraction unit and the cross-section layout unit are configured as follows: ① Extract elevation points along the cross-section line according to the preset step size; ② The Douglas-Peucker and Visvalingam-Whyatt algorithms were used to extract elevation feature points; The 3D interactive unit connects to the feature extraction unit and is configured to overlay the DOM and DEM to generate 3D terrain, supporting cross-section point editing. The output unit connects to the 3D interactive unit and is configured to automatically calculate mileage and export results in batches.

8. The automated production system for surveying cross-section products according to claim 7, characterized in that, The three-dimensional interactive unit includes: The display and editing sub-unit is configured to overlay automatically extracted elevation feature points and measured points on the real 3D terrain generated using DEM and DOM, and respond to user operations to add or delete cross-section points and label feature points.

9. A computer-readable storage medium storing a computer program, characterized in that, When the program is executed by a processor, it implements the steps of the method as described in any one of claims 1-6.

Images