A method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters

Abstract

This invention relates to the field of terrestrial lidar point cloud data processing technology, specifically a method for normalizing terrestrial lidar point clouds in bamboo forests after snow disasters. It primarily addresses the technical problem of existing normalization techniques exhibiting severe discontinuities such as "jagged edges" and "faults" in the normalized point cloud when processing bamboo points with morphological variations (bending, breaking, lodging, etc.) caused by snow disasters. This is due to incomplete extraction of ground points and abrupt changes in bamboo morphology. The method of this invention first calculates the initial relative elevation by constructing an elevation benchmark model; then, it employs a layered sliding window median smoothing technique to perform local elevation correction of the point cloud along the elevation direction using overlapping windows; for missed points not covered by the windows, global K-nearest neighbor median smoothing is used as a fallback; finally, based on the smoothing iterations, the results are fused and shifted to zero to output the normalized point cloud.

Description

Technical Field

[0001] This invention relates to the field of ground-based lidar point cloud data processing technology, and in particular to a method for normalizing ground-based lidar point clouds in bamboo forests after a snow disaster. Background Technology

[0002] Ground-based lidar technology has become an important tool in forestry resource surveys, providing a foundation for tree parameter inversion and morphological analysis by acquiring high-precision three-dimensional point cloud data. In the point cloud data processing workflow, point cloud normalization is a key preprocessing step, aiming to eliminate the influence of terrain undulations on elevation information and convert point cloud elevation values ​​to a relative elevation system based on the ground. Existing technologies typically employ algorithms based on irregular triangular networks or fabric-based analog filtering to generate digital elevation models, thereby achieving point cloud elevation normalization.

[0003] These traditional normalization methods have significant limitations when dealing with disaster-stricken forest scenarios. Especially after natural disasters such as snowstorms, bamboo forests experience large-scale morphological variations, including bending, breaking, lodging, and uprooting of bamboo stalks, resulting in exceptionally complex understory topography. In such scenarios, traditional methods often lead to incomplete ground point extraction due to vegetation cover and abrupt topographic changes, resulting in biased elevation models.

[0004] Existing methods only focus on eliminating the impact of terrain, failing to consider the damage to the point cloud structure caused by abrupt changes in the morphology of the damaged bamboo stalks. When traditional normalization is applied directly to point clouds containing bent or broken bamboo stalks, it exacerbates the structural distortion of the point cloud data, producing obvious point cloud discontinuities at the points of abrupt changes in bamboo stalk morphology, manifested as jagged distributions and elevation faults. This severely disrupts the morphological continuity of the bamboo stalks, posing significant challenges to subsequent analysis. Summary of the Invention

[0005] The purpose of this invention is to provide a method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, in order to solve the problems in the background art.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters includes the following steps:

[0008] Step 1: Extract the entire ground using Trimble RealWorks software. Employ an improved progressive densification triangular mesh filtering algorithm to extract ground points from the original point cloud, generating a classified ground point cloud.

[0009] Step 2: Using the two-dimensional planar coordinates of the classified ground point cloud, construct a spatial index model to generate an elevation benchmark model with efficient query capabilities;

[0010] Step 3: Using the elevation benchmark model, perform initial relative elevation calculations on the complete sample point cloud to obtain the initial relative elevation values ​​for each point;

[0011] Step 4: According to the preset layer height and sliding step size, perform layered sliding window median smoothing on the initial relative elevation value to generate a cumulative point cloud elevation value that includes the smoothed elevation value and the number of smoothing times;

[0012] Step 5: Based on the two-dimensional plane coordinates, perform global K-nearest neighbor median smoothing on the uncovered points after the layered sliding window median smoothing process, and output the final relative elevation values ​​of the missing points;

[0013] Step 6: Based on the number of smoothing cycles in the cumulative elevation values, perform result fusion and translation-to-zeroing processing on the smoothed elevation values, and output the normalized final point cloud dataset.

[0014] Preferably, step one specifically includes:

[0015] The entire ground region is extracted from the raw point cloud using preprocessing software.

[0016] An improved progressive encryption triangular mesh filtering algorithm is used to refine accurate ground points from the extracted ground region.

[0017] Load the three-dimensional spatial coordinate information and reflection intensity information of the complete sample location cloud;

[0018] The improved progressive encrypted triangular mesh filtering algorithm enhances terrain fitting accuracy by iteratively optimizing ground seed points.

[0019] Preferably, step two includes:

[0020] The spatial index module is constructed using a KD-tree data structure; the KD-tree data structure organizes ground point data by recursively partitioning the two-dimensional coordinate space; and the KD-tree data structure is used to achieve fast nearest neighbor queries, providing accelerated support for elevation benchmark queries.

[0021] Preferably, in step four, the median smoothing of the layered sliding window specifically includes:

[0022] Multiple sliding windows with fixed floor height intervals are set along the relative elevation direction;

[0023] Set the sliding step size to half the layer height interval to form a sequence of continuously overlapping sliding windows;

[0024] For each set of points falling within a sliding window, perform a K-nearest neighbor search based on its two-dimensional planar coordinates;

[0025] Calculate the median statistic of the relative elevation values ​​of the K nearest neighbors for each point;

[0026] Update the relative elevation of the center point using the median statistic;

[0027] Record the cumulative elevation value of each point after smoothing and the total number of times it has been smoothed.

[0028] Preferably, the floor height interval is in the range of 0.01 meters to 0.2 meters;

[0029] The layer height interval is dynamically adjusted based on the point cloud density and terrain complexity; a smaller layer height interval value is used for areas with complex terrain, and a larger layer height interval value is used for areas with gentle terrain.

[0030] Preferably, the K value in the K-nearest neighbor search is dynamically adjusted according to the local point cloud density;

[0031] The K value is set to a range of 5 to 25; a larger K value is used in dense point cloud regions and a smaller K value is used in sparse point cloud regions; the optimal K value parameter is automatically determined by local point cloud density estimation.

[0032] Preferably, the formula for calculating the median statistic is: ,in, It represents the relative elevation values ​​of a point and its K nearest neighbors, and Median is the median function; the median statistic is robust to abnormal elevation values ​​and can effectively maintain the original morphological characteristics of the point cloud.

[0033] Preferably, in step five, the global K-nearest neighbor median smoothing uses the same K-value parameter as the hierarchical sliding window median smoothing;

[0034] The same median statistical algorithm is used to ensure the consistency of the processing results; the global K-nearest neighbor median smoothing is specifically used to handle missing points that are not covered by any sliding window; global search ensures that 100% of the point cloud data has been smoothed.

[0035] Preferably, in step six, the result fusion and translation to zero specifically includes:

[0036] For a point that has been smoothed at least once, divide its cumulative elevation value by the number of times it has been smoothed to obtain the average smoothed elevation value.

[0037] The final relative elevation values ​​of all points are shifted as a whole to make the minimum elevation value in the dataset zero.

[0038] The translation process is achieved using the following formula: ,in, This is the final processed elevation value of the point. The minimum elevation value among all points. This is the final elevation value after normalization.

[0039] Preferably, a post-processing step is included after step six:

[0040] Perform statistical filtering on the normalized final point cloud dataset to remove possible outlier noise points;

[0041] Filter threshold parameters are set based on statistical analysis of point cloud density;

[0042] The processed point cloud data is output as a general lidar point cloud format file.

[0043] The format files include LAS and XYZ formats, which are used for subsequent point cloud analysis and feature extraction.

[0044] The beneficial effects of this invention are:

[0045] This invention employs an innovative dual smoothing mechanism of layered sliding windows and global fallback, significantly improving the data discontinuity problem in point clouds caused by morphological variations in bamboo stalks, such as bending and breakage, resulting from snow disasters. While eliminating the influence of terrain, this method effectively smooths out jagged fluctuations and fault phenomena, allowing the normalized point cloud to more realistically and coherently reflect the actual three-dimensional morphology of the affected bamboo, thus laying a reliable data foundation for subsequent accurate parameter inversion.

[0046] The elevation smoothing strategy based on two-dimensional spatial proximity employed in this invention reduces the absolute dependence on the completeness of ground point extraction. Even in situations where complex terrain and harsh forest environments lead to partial omissions or biases in ground point extraction, this method can still correct elevation anomalies through its unique smoothing algorithm, obtaining stable and reliable normalized results. This feature makes it particularly suitable for special application scenarios where forest stand structure is disordered after snow disasters and ground points are difficult to obtain accurately. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating a ground-based lidar point cloud normalization processing method for bamboo forests after a snow disaster, according to the present invention.

[0048] Figure 2 This invention provides a method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, which extracts a schematic diagram of the entire ground.

[0049] Figure 3 This is a schematic diagram illustrating the method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, as described in this invention, to trim the ground size.

[0050] Figure 4This is a schematic diagram of ground point extraction in a ground-based lidar point cloud normalization processing method for bamboo forests after a snow disaster, according to the present invention.

[0051] Figure 5 This is a schematic diagram of point cloud normalization in a ground-based lidar point cloud normalization processing method for bamboo forests after a snow disaster, according to the present invention. Detailed Implementation

[0052] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0053] Reference Figure 1-5 As shown in this embodiment, a method for normalizing ground-based lidar point clouds applied to bamboo forests after a snow disaster includes the following steps:

[0054] Step 1: Extract the entire ground using Trimble RealWorks software. Employ an improved progressive densification triangular mesh filtering algorithm to extract ground points from the original point cloud, generating a classified ground point cloud.

[0055] Step 2: Using the two-dimensional planar coordinates of the classified ground point cloud, construct a spatial index model to generate an elevation benchmark model with efficient query capabilities;

[0056] Step 3: Using the elevation benchmark model, perform initial relative elevation calculations on the complete sample point cloud to obtain the initial relative elevation values ​​for each point;

[0057] Step 4: According to the preset layer height and sliding step size, perform layered sliding window median smoothing on the initial relative elevation value to generate a cumulative point cloud elevation value that includes the smoothed elevation value and the number of smoothing times;

[0058] Step 5: Based on the two-dimensional plane coordinates, perform global K-nearest neighbor median smoothing on the uncovered points after the layered sliding window median smoothing process, and output the final relative elevation values ​​of the missing points;

[0059] Step 6: Based on the number of smoothing cycles in the cumulative elevation values, perform result fusion and translation-to-zeroing processing on the smoothed elevation values, and output the normalized final point cloud dataset.

[0060] In this embodiment of the invention, the entire ground surface is extracted using Trimble RealWorks software. An improved progressively encrypted triangular mesh filtering algorithm is employed to extract ground points from the original point cloud, generating a classified ground point cloud. Specific implementation methods include:

[0061] In the data preprocessing and initial ground point extraction stage, the raw point cloud data acquired by the ground-based lidar was first preprocessed to remove obvious noise and outliers. Then, using professional tools provided by Trimble RealWorks software, based on region growing algorithms and morphological filtering principles, the complete ground region was quickly extracted from the raw point cloud. This step utilizes the software's built-in advanced algorithms to effectively identify large-scale terrain features, providing reliable foundational data for subsequent processing.

[0062] After obtaining the ground region extracted by the software, a gridded seed point selection process is implemented. The extracted ground region is divided into regular grid cells, and the lowest point within each grid is calculated as the initial seed point. These seed points form the basis for constructing the initial irregular triangular mesh. The initial terrain model established through the grid method can reflect the approximate outline of the terrain, laying the foundation for subsequent iterative refinement processing. Although the initial triangular mesh model has limited accuracy, it can effectively eliminate most interference from non-ground objects, providing a robust starting point for the algorithm.

[0063] Entering the core iterative encryption filtering stage, based on the initial triangulation model, all data points not classified as ground points are traversed. The algorithm calculates the spatial geometric relationship between each point and the nearest triangle in the triangulation, employing a dual judgment criterion: evaluating both the perpendicular distance from the point to the triangle plane and analyzing the spatial angle between the point and the triangle's normal vector. Considering the unique environment of bamboo forests after a snow disaster, relatively lenient initial angle and distance thresholds are set. Only when a data point simultaneously meets both the distance and angle threshold conditions is it identified as a candidate ground point and added to the ground point set, while the local irregular triangulation model is dynamically updated.

[0064] Dynamic threshold optimization and terrain fitting are implemented. After each complete iteration, the angle and distance thresholds are simultaneously reduced according to a preset shrinkage coefficient. This dynamic adjustment mechanism allows the algorithm to adapt to complex terrain undulations caused by snow accumulation and fallen bamboo poles in the early stages, accurately capturing macroscopic terrain features. In later stages, stricter criteria are used to screen for real ground points in minute details, effectively avoiding misclassifying bamboo poles or residual vegetation close to the ground as ground. Through this gradual strategy from coarse to fine, the algorithm can adaptively fit an accurate surface model.

[0065] After iterative encryption, post-processing and result optimization are performed. The initially acquired ground point set undergoes manual inspection and morphological filtering to remove any potentially remaining non-ground points. The optimized high-precision ground point cloud is then integrated with the non-ground points in the original point cloud to generate a complete sample point cloud dataset containing accurate classifications. The high-quality ground point cloud generated in this stage serves as the direct data source for constructing the elevation benchmark model; its accuracy has a decisive impact on the final effect of the entire normalization process and is a crucial foundation for the successful implementation of this method.

[0066] In this embodiment of the invention, a spatial index model is constructed using the two-dimensional planar coordinates of the classified ground point cloud to generate an elevation benchmark model with efficient query capabilities. Specific implementation methods include:

[0067] Data structure initialization and spatial partitioning are performed. The classified ground point cloud data is loaded into the processing system, and the two-dimensional planar coordinates (x, y) and corresponding elevation value z of each ground point are extracted. A KD-tree spatial index structure is constructed based on these two-dimensional coordinates. The construction process employs a recursive partitioning strategy, selecting the dimension with the largest data variance as the partitioning dimension and the median point as the partition point, ensuring that the generated binary tree remains balanced. Each tree node stores statistical information about the point set within its corresponding spatial range, including the number of points, coordinate range, and elevation distribution characteristics. This hierarchical spatial partitioning establishes an optimized data structure foundation for subsequent efficient spatial queries.

[0068] The KD-tree is constructed and optimized. The construction process starts from the root node and recursively divides the 2D space into two non-overlapping sub-regions. The splitting hyperplane is perpendicular to the current splitting dimension coordinate axis and passes through the selected points. Each internal node stores the splitting dimension and splitting threshold, while leaf nodes store the set of ground points falling within that spatial partition. To improve query efficiency, a leaf node capacity threshold is set; if the number of points within a node exceeds this threshold, the splitting continues. After construction, tree structure optimization is performed, including subtree balancing and node data compression, ensuring minimal tree depth and optimized query paths. The optimized KD-tree can complete nearest neighbor search in average time complexity.

[0069] To further establish an elevation interpolation model, a KD-tree index structure is used to construct an elevation benchmark model applicable to the entire sampling area. For any query point p(x, y), the leaf node region where it is located is quickly located by traversing the KD-tree. The elevation benchmark value of this point is calculated using an inverse distance weighted interpolation algorithm. The interpolation formula is: The weight , This indicates the distance from query point p to nearby ground points. Euclidean distance, It is the weight value of the i-th neighboring point. It is the elevation value of the i-th nearest neighbor point. A KD-tree is used to efficiently retrieve the k nearest neighbor ground points around the query point, and a weighted average of the elevation values ​​of these neighbor points is performed to generate a continuous and smooth elevation datum.

[0070] To improve query efficiency of the elevation datum model, several optimization strategies were implemented based on the KD-tree structure. These included pre-calculating elevation values ​​for commonly used query grids and establishing a query caching mechanism; implementing batch query processing to reduce tree traversal overhead; and employing an approximate nearest neighbor search algorithm to improve query speed while maintaining accuracy. For large-scale point cloud data processing, a parallel query architecture was designed, dividing the query region into multiple sub-regions and utilizing multi-threading technology to process multiple query requests simultaneously. These optimizations significantly improved the query performance of the elevation datum model when dealing with massive amounts of point cloud data.

[0071] Model validation and accuracy assessment were conducted, using cross-validation to evaluate the accuracy and reliability of the elevation benchmark model. Randomly selected ground points with known elevations were used as validation points to compare the differences between the model-predicted elevation values ​​and the actual measured values. The root mean square error was calculated. Evaluate the model accuracy, where n is the number of validation points. The model-predicted elevation value of the i-th point. : The actual measured elevation value of the i-th point. Simultaneously, the spatial distribution characteristics of the error are analyzed, and local model optimization is performed for areas with large errors, including increasing local sampling density and adjusting interpolation parameters. The validated elevation benchmark model will serve as the reference for subsequent point cloud normalization processing.

[0072] In this embodiment of the invention, the initial relative elevation of the complete sample point cloud is calculated using the elevation benchmark model to obtain the initial relative elevation value of each point. Specific implementation methods include:

[0073] Data preparation and model loading are performed by loading complete sample point cloud data, including all ground and non-ground points, into the processing system. Simultaneously, an elevation datum model is loaded, constructed based on a KD-tree spatial index structure and an inverse distance weighted interpolation method. The system establishes the coordinate correspondence between the point cloud data and the elevation datum model, ensuring that both are in the same plane coordinate system and under the same elevation datum. The relative elevation calculation result storage structure is initialized, allocating storage space for each point to record its initial relative elevation value and related calculation status.

[0074] Perform point cloud traversal and benchmark elevation query, processing each point in the complete sample point cloud using a sequential traversal approach. For the currently processed point... Extract its two-dimensional plane coordinates As input for the query, the elevation datum value corresponding to this plane location is obtained through the query interface provided by the elevation datum model. The query process utilizes a KD-tree spatial index to achieve efficient nearest neighbor search, locates the spatial partition containing the point, retrieves nearby ground points, and calculates accurate benchmark elevation values ​​based on an inverse distance weighted algorithm.

[0075] Perform relative elevation calculations and record the results to obtain the points. The benchmark elevation value Then, calculate its initial relative elevation value. The calculation formula is: ,in Point The original absolute elevation value, This represents the elevation datum value for the corresponding location. This is the calculated initial relative elevation value. This calculation process converts the absolute elevation of a point into a relative elevation based on the local terrain, eliminating the influence of terrain undulations on the point cloud elevation value. The system then displays the calculation results. The point identification information is stored in the output dataset.

[0076] Outlier detection and result verification are performed, with data quality control implemented concurrently during the calculation process. A reasonable threshold range for relative elevation values ​​is set, and outliers exceeding this threshold are marked and reviewed. Outlier review includes checking the accuracy of the original point cloud coordinates, verifying the reliability of the elevation benchmark model query results, and confirming the correctness of the coordinate transformation parameters. For confirmed outliers, their abnormal states are recorded, allowing for selective exclusion and repair using interpolation methods in subsequent processing stages.

[0077] After processing all points, the calculation results are integrated and output. The initial relative elevation calculation results are then combined to form a complete dataset. This dataset retains the planar coordinates and intensity information of the original point cloud, while also including the calculated initial relative elevation values. The output data is stored in a standardized format to ensure compatibility with subsequent processing modules. A processing report is generated, summarizing key indicators from the calculation process, including the number of processed points, the number of outliers, and calculation accuracy assessments, providing a quality reference for subsequent processing steps.

[0078] In this embodiment of the invention, the initial relative elevation value is subjected to layered sliding window median smoothing according to a preset layer height and sliding step size, generating a cumulative point cloud elevation value that includes the smoothed elevation value and the number of smoothing times. Specific implementation methods include:

[0079] Perform parameter initialization and data structure preparation, and set the core parameters for layered sliding window processing: layer height interval. With sliding step size The sliding step size is half of the floor height interval, i.e. Establish a storage structure for cumulative point cloud elevation values, allocating storage space for each point to record the cumulative smoothed elevation value and the number of smoothing iterations. Initialize a sliding window sequence, starting from the minimum relative elevation value along the relative elevation direction, and sequentially create sliding windows covering different elevation intervals in increments of the sliding step size. Each sliding window records the elevation range it covers. and the index of the point set it contains.

[0080] Construct a sliding window sequence and assign point sets. Based on the distribution range of the initial relative elevation values, calculate the required number of sliding windows N: ,in and Let represent the maximum and minimum initial relative elevation values ​​in the point cloud, respectively. The elevation range of each sliding window is generated sequentially, with the coverage area of ​​the i-th window being... Iterate through each point in the point cloud, determine the sliding window it falls into based on its initial relative elevation value, and add the point index to the point set of the corresponding window. Establish a many-to-many mapping relationship between points and windows, and record which sliding windows cover each point.

[0081] Perform K-nearest neighbor search and median calculation on the point set within the window. For each point in the sliding window, perform the K-nearest neighbor search point by point. Based on its two-dimensional plane coordinates Search for the K nearest neighbors in the current sliding window's point set. Calculate the median statistic of the relative elevation values ​​of these nearest neighbors: ,in arrive This represents the relative elevation values ​​of the K nearest neighbors, with Median being the median function. The value of K is dynamically determined based on the point cloud density; a larger K value is used for regions with high density, and a smaller K value is used for regions with low density, ensuring statistical stability.

[0082] Implement elevation value updates and cumulative records, and calculate the median statistic. As a point Smoothing results in the current window. Update point. The cumulative elevation value is calculated by adding the current value to its original cumulative value. At the same time, add some points. A smoothing counter is used. A point status marking mechanism is established to record the processing status of each point in each window, preventing duplicate counting. For each sliding window, the number of processed points and processing progress are recorded to ensure processing continuity between windows.

[0083] To address window boundary effects and point set overlap, since overlapping areas exist between sliding windows, some points may be covered by multiple windows. A strategy for handling overlapping points is designed to ensure that these points can participate independently in the smoothing calculation within each covering window. The point set overlap rate at window boundaries is calculated to verify the rationality of the overlap settings. For points located at the window edges, the completeness of their K-nearest neighbor search is specifically checked to ensure that nearest neighbor points are not missed due to window segmentation.

[0084] Implement quality control and anomaly handling, monitoring data quality in real time during the smoothing process. Detect anomalies in median calculation, including insufficient nearest neighbors and abnormally dispersed elevation values. For cases with excessively large variance in the elevation values ​​of the K nearest neighbors, activate alternative processing schemes, such as mean filtering or expanding the search range. Record the number of outliers encountered during processing and the processing results, generating a quality control report.

[0085] After completing the result integration and data output, integrate the cumulative elevation values ​​of the point cloud after all sliding windows have been processed. Generate a complete dataset containing the cumulative smoothed elevation value and the number of smoothing iterations for each point. Verify the integrity of the data processing to ensure that all points falling within the sliding window have been correctly processed. Output a detailed processing report for layered sliding window median smoothing, including window parameters, statistics on the number of processed points, quality assessment indicators, and other information. Provide a standardized data interface for subsequent processing steps.

[0086] To optimize processing efficiency and resource management, a parallel processing strategy is adopted for large-scale point cloud data. Sliding windows are distributed across multiple processing threads, processing point sets within multiple windows simultaneously. An efficient memory management mechanism is designed to load window point set data on demand, reducing memory consumption. A processing progress monitoring system is established to display the processing status of each window and the overall processing progress in real time. The K-nearest neighbor search algorithm is optimized, utilizing a spatial index structure to improve search efficiency, ensuring the completion of the massive point cloud processing task within the specified time.

[0087] Parameter sensitivity analysis and effect verification were conducted, and the floor height spacing was analyzed through experiments. The influence of the K-value on the smoothing effect is investigated. Typical regions are selected to test the processing results under different parameter combinations, evaluating the impact of parameter changes on point cloud smoothness and detail preservation. Parameter optimization selection criteria are established to provide parameter configuration suggestions for different point cloud features. The elevation distribution characteristics of point clouds before and after smoothing are compared to verify the effectiveness of layered sliding window median smoothing in eliminating noise and maintaining morphological continuity.

[0088] In this embodiment of the invention, based on two-dimensional planar coordinates, global K-nearest neighbor median smoothing is performed on the uncovered points after layered sliding window median smoothing, and the final relative elevation value of the missing points is output. Specific implementation methods include:

[0089] An uncovered point identification mechanism is established. The system traverses each data point in the point cloud dataset and identifies all points not covered by the layered sliding window by querying their smoothing count record. These uncovered points are mainly distributed in: elevation gaps in the sliding window sequence, areas of bamboo morphological variation, and areas where morphological variations intersect or mix with surrounding bamboo. The system establishes an independent set of uncovered points, recording the spatial location information and initial relative elevation value of each point, while retaining its index mapping relationship with the original point cloud data.

[0090] A global processing environment is constructed, and a spatial index structure is rebuilt based on the two-dimensional planar coordinates of the complete sample point cloud. This index covers all data points, both those that have undergone hierarchical sliding window processing and those that have not. A balanced KD-tree algorithm is used to organize the spatial data, ensuring efficient nearest neighbor query performance when processing massive point clouds. The index structure includes the spatial coordinates, processing status flags, and elevation information for each point, providing data support for subsequent global searches.

[0091] A global nearest neighbor search and elevation calculation are implemented. For each uncovered point, the system performs a K-nearest neighbor search in its two-dimensional plane coordinate space. The search range is not limited by elevation layering; it finds the K spatially closest points from the entire point cloud data. The relative elevation values ​​of these nearest neighbors are collected, sorted in ascending order, and the median is taken as the final relative elevation value of the uncovered point. The calculation process employs robust statistical methods to ensure reasonably smooth results even when the point cloud density is uneven.

[0092] The system performs data processing and status updates, directly assigning the calculated median elevation result as the final relative elevation value for the uncovered point. Simultaneously, it updates the processing status flag for that point, marking it as having completed smoothing. The system maintains a dedicated processing log, recording the original elevation, processed elevation, and distribution characteristics of its nearest neighbors for each uncovered point, facilitating subsequent quality traceability and effectiveness evaluation.

[0093] The system performs quality verification and anomaly handling. It automatically checks the global smoothing results to verify the elevation continuity between uncovered points and surrounding point clouds after processing. For anomalies encountered during processing, such as insufficient neighboring points or abnormally discrete elevation values, the system activates a backup processing mechanism. By gradually expanding the search range or introducing elevation constraints, the system ensures that all uncovered points receive reasonable elevation assignments.

[0094] After completing data integration and output preparation, and processing all uncovered points, the system integrates the global smoothing results with the hierarchical sliding window processing results to form a unified point cloud elevation dataset. The completeness and consistency of the dataset are verified to ensure that each point has a definite final relative elevation value. Output preparation includes data format standardization, metadata generation, and processing report compilation, providing complete and reliable data input for subsequent processing stages.

[0095] In this embodiment of the invention, based on the number of smoothing iterations in the cumulative elevation value, the smoothed elevation value is subjected to result fusion and translation-to-zeroing processing to output a normalized final point cloud dataset. Specific implementation methods include:

[0096] The system reads the cumulative elevation dataset of point clouds after layered sliding window smoothing and global smoothing. Data integrity is verified to ensure that each point contains a record of the cumulative smoothed elevation value and the number of smoothing cycles. A point cloud state classification mechanism is established, dividing points into three processing categories: points that have undergone multiple smoothing cycles, points that have undergone only one smoothing cycle, and points that have undergone global smoothing but not layered smoothing. A corresponding result fusion strategy is designed for each category.

[0097] For points with a smoothing order greater than zero, calculate their average smoothed elevation value. The formula is: ,in This represents the elevation value obtained after the j-th smoothing process. This indicates the total number of smoothing iterations at that point. This represents the calculated average smoothed elevation value. For points that have undergone only a single smoothing process, the average elevation value is equal to the elevation value obtained during that smoothing process. A quality control mechanism for the average elevation value is established to perform range verification and outlier detection on the calculation results.

[0098] Perform translation and zeroing processing, scan the average smoothed elevation values ​​of all points, and determine the minimum elevation value in the dataset. Perform the translation transformation formula: This transformation ensures that all elevation values ​​in the normalized point cloud dataset are non-negative, and the minimum elevation value is zero. The mathematical correctness of the translation transformation is verified, confirming that the transformation process does not change the spatial relationships and morphological features of the point cloud. The translation transformation parameters are recorded, including... The numerical value and the transformation timestamp.

[0099] Data quality assessment and post-processing were performed, including a comprehensive quality check on the normalized point cloud dataset. The distribution characteristics of elevation values ​​were analyzed to verify their conformity to expected statistical patterns. Outliers were detected and addressed, including negative elevation values, abnormally large values, and elevation values ​​that clearly violated spatial continuity. Data consistency verification was conducted to ensure the correct correspondence between planar coordinates and elevation values.

[0100] Complete data formatting and output, converting the processed normalized point cloud data into a standard format. The data output includes the three-dimensional coordinate information of each point, where the planar coordinates retain their original values, and the elevation coordinates are normalized. The system outputs processing metadata, including processing parameters, quality assessment results, and transformation records. A data version management mechanism is established to ensure the traceability of output data and the processing process. Multiple data interfaces are provided for subsequent applications to support compatibility requirements across different platforms.

[0101] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0102] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, characterized in that, Includes the following steps: Step 1: Using an improved progressive densification triangular mesh filtering algorithm, ground point extraction is performed on the original point cloud to generate a classified ground point cloud; Step 2: Using the two-dimensional planar coordinates of the classified ground point cloud, construct a spatial index model to generate an elevation benchmark model with efficient query capabilities; Step 3: Using the elevation benchmark model, perform initial relative elevation calculations on the complete sample point cloud to obtain the initial relative elevation values ​​for each point; Step 4: According to the preset layer height and sliding step size, perform layered sliding window median smoothing on the initial relative elevation value to generate a cumulative point cloud elevation value that includes the smoothed elevation value and the number of smoothing times; Step 5: Based on the two-dimensional plane coordinates, perform global K-nearest neighbor median smoothing on the uncovered points after the layered sliding window median smoothing process, and output the final relative elevation values ​​of the missing points; Step 6: Based on the number of smoothing cycles in the cumulative elevation values, perform result fusion and translation-to-zeroing processing on the smoothed elevation values, and output the normalized final point cloud dataset; In step four, the layered sliding window median smoothing specifically includes: Multiple sliding windows with fixed floor height intervals are set along the relative elevation direction; Set the sliding step size to half the layer height interval to form a sequence of continuously overlapping sliding windows; For each set of points falling within a sliding window, perform a K-nearest neighbor search based on its two-dimensional planar coordinates; Calculate the median statistic of the relative elevation values ​​of the K nearest neighbors for each point; Update the relative elevation of the center point using the median statistic; Record the cumulative elevation value of each point after smoothing and the total number of times it has been smoothed; The value of the floor height interval ranges from 0.01 meters to 0.2 meters; The layer height interval is dynamically adjusted based on the point cloud density and terrain complexity; a smaller layer height interval value is used for areas with complex terrain, and a larger layer height interval value is used for areas with gentle terrain. The K value in the K-nearest neighbor search is dynamically adjusted according to the local point cloud density; The value of K is set to range from 5 to 25; The formula for calculating the median statistic is: ,in, It represents the relative elevation values ​​of a point and its K nearest neighbors, where Median is the median function; In step five, the global K-nearest neighbor median smoothing uses the same K-value parameter as the hierarchical sliding window median smoothing. The global K-nearest neighbor median smoothing is specifically designed to handle missing points that are not covered by any sliding window; a global search ensures that 100% of the point cloud data has been smoothed.

2. The method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, as described in claim 1, is characterized in that... Step one specifically includes: The entire ground region is extracted from the raw point cloud using preprocessing software. An improved progressive encryption triangular mesh filtering algorithm is used to refine accurate ground points from the extracted ground region. Load the three-dimensional spatial coordinate information and reflection intensity information of the complete sample location cloud; The improved progressive encrypted triangular mesh filtering algorithm enhances terrain fitting accuracy by iteratively optimizing ground seed points.

3. The method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, as described in claim 1, is characterized in that... Step two includes: The spatial index module is constructed using a KD-tree data structure; the KD-tree data structure organizes ground point data by recursively partitioning the two-dimensional coordinate space; and the KD-tree data structure is used to achieve fast nearest neighbor queries, providing accelerated support for elevation benchmark queries.

4. The method for normalizing ground-based lidar point clouds in bamboo forests after a snow disaster, as described in claim 1, is characterized in that... In step six, the result fusion and translation to zero specifically includes: For a point that has been smoothed at least once, divide its cumulative elevation value by the number of times it has been smoothed to obtain the average smoothed elevation value. The final relative elevation values ​​of all points are shifted as a whole to make the minimum elevation value in the dataset zero. The translation process is achieved using the following formula: ,in, This is the final processed elevation value of the point. The minimum elevation value among all points. This is the final elevation value after normalization.

5. The method for normalizing ground-based lidar point clouds in bamboo forests after snow disasters, as described in claim 1, is characterized in that... Following step six, post-processing steps are also included: Perform statistical filtering on the normalized final point cloud dataset to remove possible outlier noise points; Filter threshold parameters are set based on statistical analysis of point cloud density; The processed point cloud data is output as a general lidar point cloud format file. The format files include LAS and XYZ formats, which are used for subsequent point cloud analysis and feature extraction.

Images