Road surface texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation

Abstract

The invention belongs to the technical field of road surface detection and point cloud intelligent reconstruction, and particularly relates to a road surface texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation, which comprises the following steps: acquiring road surface point cloud data, and preprocessing to obtain an effective analysis area and effective road surface point cloud data; determining a missing rate based on the effective analysis area and the effective pavement point cloud data; determining an interpolation algorithm according to the missing rate, and repairing the point cloud data to obtain complete point cloud data; and sequentially constructing a three-dimensional pavement model based on the complete point cloud data, calculating multi-scale texture features, performing texture optimization and consistency correction, and finally outputting pavement curved surface data. According to the method, a multi-algorithm fusion data interpolation strategy of missing rate gating is provided, the reconstruction efficiency and accuracy are ensured by adopting an inverse distance weighted KNN algorithm in low missing, and the K value is adaptively optimized by adopting a texture-driven EKNN algorithm in high missing, so that the texture reconstruction precision and robustness are improved.

Description

Technical Field

[0001] This invention belongs to the field of road surface detection and intelligent point cloud reconstruction technology, specifically involving a road surface texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation. Background Technology

[0002] Laser scanning of road surfaces to obtain high-precision 3D elevation data, construct accurate 3D models, and extract macroscopic and microscopic texture indicators has become an important means of evaluating road surface friction performance, structural safety, and durability. In recent years, vehicle-mounted 3D laser scanners have been widely used in road surface elevation measurement and texture analysis due to their non-contact, high-efficiency, and high-precision characteristics. However, in actual scanning processes, due to limitations in equipment measurement accuracy, vehicle vibration, environmental occlusion, and complex weather conditions, road surface point cloud data inevitably exhibits local clustering or significant data gaps. If these missing data are not effectively processed, they will directly affect the calculation accuracy of macroscopic and microscopic texture parameters, thereby reducing the reliability of road performance evaluation and structural analysis.

[0003] High-precision interpolation and texture reconstruction of missing data from vehicle-mounted laser point clouds has become an urgent problem to be solved. There is an urgent need for a high-precision, high-efficiency, and high-robust compensation and reconstruction technology to restore road elevation and texture information, and provide reliable data support for road maintenance decisions, performance evaluation, and structural analysis. Summary of the Invention

[0004] The purpose of this invention is to provide a road surface texture reconstruction method based on missing point cloud data interpolation by vehicle-mounted laser, which aims to overcome the problems of missing point cloud data, low repair accuracy and insufficient computational efficiency, and achieve high-precision and high-efficiency reconstruction of road surface texture features.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for road surface texture reconstruction based on interpolation of missing data from vehicle-mounted laser point clouds, comprising: Acquire road surface point cloud data and perform preprocessing to obtain the effective analysis area and effective road surface point cloud data; The missing rate is determined based on the effective analysis area and effective road surface point cloud data; The imputation algorithm is determined based on the missing rate, and the point cloud data is repaired to obtain complete point cloud data; Based on complete point cloud data, a 3D road surface model is constructed sequentially, multi-scale texture features are calculated, and texture optimization and consistency correction are performed to finally output road surface data.

[0006] Preferably, the preprocessing includes: The central rectangular region is extracted from the collected road surface point cloud data as the effective analysis area. The point cloud data within this area is then subjected to noise reduction and smoothing to eliminate outliers, thus obtaining effective road surface point cloud data.

[0007] Preferably, determining the missing rate based on the effective analysis area and effective road surface point cloud data includes: The effective analysis area is divided into grids, and a regular two-dimensional coordinate grid matrix is ​​established. Based on the spatial coordinates of the point cloud, each effective point is mapped to the corresponding two-dimensional coordinate grid matrix. Identify missing points within all two-dimensional coordinate raster matrices and count the number of missing points; The missing rate is calculated based on the number of missing points and the effective road surface point cloud data.

[0008] Preferably, the step of determining the imputation algorithm based on the missing rate and performing point cloud data repair to obtain complete point cloud data includes: When the missing rate is ≤10%, point cloud data repair is performed using the inverse distance weighted KNN algorithm to obtain complete point cloud data; When the missing rate is greater than 10%, the texture-driven EKNN algorithm is used to repair the point cloud data and obtain complete point cloud data.

[0009] Preferably, the point cloud data repair using the inverse distance weighted KNN algorithm includes: For each missing point, identify its set of candidate neighbor points in a two-dimensional raster matrix, and calculate the similarity between the missing point and each candidate neighbor point; Based on similarity ranking, select the K most similar neighbor points as the nearest neighbor set; The elevation values ​​of the missing points are obtained by weighted averaging of the elevations of the selected K neighboring points, which is the estimated missing point. The estimated missing points are included in the neighborhood search, and the above steps are repeated iteratively to imputate until all missing points are estimated. Perform local planar or bilinear smoothing on the interpolated point cloud.

[0010] Preferably, the point cloud data inpainting using the texture-driven EKNN algorithm includes: The observed valid points in the point cloud dataset are used as the neighborhood search sample set, and the missing points are used as the target to be estimated. Initialize the K value, setting the initial value K=3, and allowing a search range of 3~10; The optimal value of parameter K is obtained by optimizing the discrete grid search method. The missing points are finally imputed using the optimal K value to generate complete point cloud data; Perform local planar or bilinear smoothing on the complete point cloud data.

[0011] Preferably, the process of sequentially constructing a 3D road surface model based on complete point cloud data, calculating multi-scale texture features, performing texture optimization and consistency correction, and finally outputting road surface data includes: Based on the complete point cloud data, several local regions are obtained by dividing the data into blocks according to a fixed grid. For each local region, a continuous three-dimensional road surface model is established by combining local quadratic surface fitting and triangular mesh generation. Based on a three-dimensional road surface model, macroscopic and microscopic texture feature parameters for evaluating road surface function are calculated; Based on macroscopic and microscopic texture feature parameters, texture optimization and consistency correction are performed on the three-dimensional road surface model to obtain optimized road surface data. The optimized road surface data is output as a 3D model format.

[0012] In a second aspect, the present invention provides a road surface texture reconstruction system based on interpolation of missing data from vehicle-mounted laser point clouds, comprising: The acquisition unit is used to acquire road surface point cloud data and perform preprocessing to obtain the effective analysis area and effective road surface point cloud data. A calculation unit is used to determine the missing rate based on the effective analysis area and effective road surface point cloud data; The interpolation unit is used to determine the interpolation algorithm based on the missing rate and to repair the point cloud data to obtain complete point cloud data. The reconstruction unit is used to sequentially construct a 3D road surface model based on complete point cloud data, calculate multi-scale texture features, perform texture optimization and consistency correction, and finally output road surface data.

[0013] In a third aspect, the present invention provides an electronic device, characterized in that it includes a processor and a memory, wherein the processor is configured to execute a computer program stored in the memory to implement the road surface texture reconstruction method based on missing data interpolation of vehicle-mounted laser point cloud as described in any one of the preceding claims.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores at least one instruction, which, when executed by a processor, implements the road surface texture reconstruction method based on missing data interpolation of vehicle-mounted laser point cloud as described in any one of the preceding claims.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes a multi-algorithm fusion data interpolation strategy with missing rate gating. When missing data is low, an inverse distance weighted KNN algorithm is used to ensure reconstruction efficiency and accuracy. When missing data is high, a texture-driven EKNN algorithm is used to adaptively optimize the K value, improving texture reconstruction accuracy and robustness. Furthermore, this invention achieves global continuous reconstruction through local quadratic surface + TIN stitching, combined with texture consistency correction, reducing MTD / Ssk / Sku bias under different missing data states, improving the reconstruction quality of road surface texture features, and ensuring the accuracy and reliability of road surface detection and performance evaluation. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention. Figure 2 This is a flowchart of the road texture reconstruction method based on missing data interpolation of vehicle-mounted laser point cloud according to an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the method for characterizing and calculating road surface texture indicators according to an embodiment of the present invention; Figure 4 This is a comparison chart showing the reconstruction accuracy of the MTD index by different interpolation algorithms in this embodiment of the invention; Figure 5 This is a comparison chart showing the reconstruction accuracy of the Ssk index by different interpolation algorithms in this embodiment of the invention; Figure 6 This is a comparison chart showing the reconstruction accuracy of the SKU index by different interpolation algorithms in this embodiment of the invention; Figure 7 The edge of the 10% missing data interpolation region in this embodiment of the invention x Axial elevation distribution diagram; Figure 8 This is a cross-sectional elevation distribution map along the y-axis of the 10% missing data interpolation area in Embodiment 1 of the present invention; Figure 9 This is a structural block diagram of a road texture reconstruction system based on missing data interpolation of vehicle-mounted laser point cloud, according to an embodiment of the present invention. Figure 10 This is a structural block diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0018] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0019] See Figure 1 This application discloses a method for road surface texture reconstruction based on interpolation of missing data from vehicle-mounted laser point clouds, including: S1: Acquire road surface point cloud data and perform preprocessing to obtain the effective analysis area and effective road surface point cloud data; S2: Determine the missing rate based on the effective analysis area and effective road surface point cloud data; S3: Determine the imputation algorithm based on the missing rate, and perform point cloud data repair to obtain complete point cloud data; S4: Based on complete point cloud data, a 3D road surface model is constructed sequentially, multi-scale texture features are calculated, and texture optimization and consistency correction are performed to finally output road surface data.

[0020] In some embodiments, the preprocessing includes: The central rectangular region is extracted from the collected road surface point cloud data as the effective analysis area. The point cloud data within this area is then subjected to noise reduction and smoothing to eliminate outliers, thus obtaining effective road surface point cloud data.

[0021] In some embodiments, determining the missing rate based on the effective analysis area and effective road surface point cloud data includes: The effective analysis area is divided into grids, and a regular two-dimensional coordinate grid matrix is ​​established. Based on the spatial coordinates of the point cloud, each effective point is mapped to the corresponding two-dimensional coordinate grid matrix. Identify missing points within all two-dimensional coordinate raster matrices and count the number of missing points; The missing rate is calculated based on the number of missing points and the effective road surface point cloud data.

[0022] In some embodiments, determining the imputation algorithm based on the missing rate and performing point cloud data repair to obtain complete point cloud data includes: When the missing rate is ≤10%, point cloud data repair is performed using the inverse distance weighted KNN algorithm to obtain complete point cloud data; When the missing rate is greater than 10%, the texture-driven EKNN algorithm is used to repair the point cloud data and obtain complete point cloud data.

[0023] In some embodiments, the point cloud data repair using the inverse distance weighted KNN algorithm includes: For each missing point, identify its set of candidate neighbor points in a two-dimensional raster matrix, and calculate the similarity between the missing point and each candidate neighbor point; Based on similarity ranking, select the K most similar neighbor points as the nearest neighbor set; The elevation values ​​of the missing points are obtained by weighted averaging of the elevations of the selected K neighboring points, which is the estimated missing point. The estimated missing points are included in the neighborhood search, and the above steps are repeated iteratively to imputate until all missing points are estimated. Perform local planar or bilinear smoothing on the interpolated point cloud.

[0024] In some embodiments, the point cloud data inpainting using the texture-driven EKNN algorithm includes: The observed valid points in the point cloud dataset are used as the neighborhood search sample set, and the missing points are used as the target to be estimated. Initialize the K value, setting the initial value K=3, and allowing a search range of 3~10; The optimal value of parameter K is obtained by optimizing the discrete grid search method. The missing points are finally imputed using the optimal K value to generate complete point cloud data; Perform local planar or bilinear smoothing on the complete point cloud data.

[0025] In some embodiments, the road texture reconstruction method based on missing data interpolation of vehicle-mounted laser point cloud as described in claim 1 is characterized in that the step of sequentially constructing a three-dimensional road surface model based on complete point cloud data, calculating multi-scale texture features, performing texture optimization and consistency correction, and outputting road surface data includes: Based on the complete point cloud data, several local regions are obtained by dividing the data into blocks according to a fixed grid. For each local region, a continuous three-dimensional road surface model is established by combining local quadratic surface fitting and triangular mesh generation. Based on a three-dimensional road surface model, macroscopic and microscopic texture feature parameters for evaluating road surface function are calculated; Based on macroscopic and microscopic texture feature parameters, texture optimization and consistency correction are performed on the three-dimensional road surface model to obtain optimized road surface data. The optimized road surface data is output as a 3D model format.

[0026] Example 1 like Figure 2 As shown, a method for road surface texture reconstruction based on missing data interpolation from vehicle-mounted laser point clouds includes the following steps: S1: Road surface point cloud data acquisition and preprocessing; S2: Point cloud data missing rate calculation; S3: Point cloud data interpolation and repair; S4: Road surface texture reconstruction and optimization.

[0027] Step S1 specifically includes: S11: Obtain road surface elevation point cloud data using a vehicle-mounted 3D laser scanner. A Gocator 2380 rack-mounted 3D laser scanner manufactured by LMI Technologies was used to obtain road surface elevation point cloud data.

[0028] To verify the texture reconstruction effect, the aforementioned device was used to collect accurate point cloud data as ground truth. By setting different data missing rates and calculating the texture reconstruction index under different missing rates, the rationality and correctness of the present invention were verified.

[0029] During data acquisition, the scanner was set up at a height of 1.5m and moved at a speed of 200mm / s to ensure the results closely approximate real-world vehicle-mounted laser scanning data. The laser emission frequency was 182Hz, the exposure value was 700μs, the lateral spacing was 0.5mm, the longitudinal spacing was 0.5mm, and the effective field of view was 800mm.

[0030] Finally, surface elevation point cloud measurement data of 3036 road surface samples were collected as a validation dataset, including 2500 asphalt pavement test specimens and 536 cement concrete slabs. The dimensions of the specimens were 300 mm long, 300 mm wide, and 50 mm high.

[0031] S12: Extract a rectangular area of ​​250 mm × 250 mm at the center of the specimen as the effective analysis area, and perform noise reduction and smoothing on the point cloud data within the area to eliminate abnormal points caused by environmental interference, equipment vibration and other factors.

[0032] Step S2 quantifies the missing rate by calculating the proportion of missing points within the valid analysis area. Specifically, this includes: S21: Divide the effective analysis area (250 mm × 250 mm) into fixed intervals Δ x = Δ y Mesh the grid to 0.5mm and establish a regular two-dimensional coordinate grid matrix G( x i , y i Based on the spatial coordinates of the point cloud ( x , y , z This maps each valid point to the corresponding grid cell, achieving a one-to-one correspondence between spatial coordinates and grid cells.

[0033] S22: Traverse the raster matrix G( xi , y i If a certain grid cell does not contain a corresponding elevation value z ( x i , y i If a point is missing, it is considered a missing point; if multiple points exist in the same grid, their average elevation value is taken as the representative value. Calculate the number of valid points. N valid Total number of grid points N total And count the number of missing points. N miss = N total - N valid .

[0034] S23: Calculate the missing rate of point cloud data according to formula (1) R miss : (1) in, R miss Indicates the missing rate (%); N miss This represents the number of missing points; N total To effectively analyze the total number of points within the region.

[0035] In step S3, different types of imputation algorithms are used to repair point cloud data for different missing rate scenarios. Specifically, when the missing rate is ≤10%, the inverse distance weighted KNN algorithm is used for imputation to balance speed and accuracy; when the missing rate is >10%, the texture-driven EKNN algorithm is used for imputation to improve robustness in high missing rate scenarios.

[0036] To verify the rationality of the proposed data imputation strategy, missing rates were manually set for the datasets at 1%, 3%, 5%, 10%, 30%, and 50%, respectively. Different data imputation algorithms were used to restore point cloud data for datasets with different missing rates. The rationality and correctness of the proposed solution were verified by comparing the road texture reconstruction results.

[0037] The data interpolation algorithms used include statistical methods: chain equation multiple interpolation algorithm (MICE); machine learning methods: MissForest algorithm, inverse distance weighted KNN algorithm, texture-driven EKNN algorithm; and deep learning methods: generative adversarial interpolation network algorithm (GAIN).

[0038] Based on the set missing rate, a missing region is generated for each sample using random aggregation. The size of the missing matrix ranges from 20×20 to 100×100, allowing overlap until the preset missing data rate is reached. It should be noted that the location of missing values ​​is randomly assigned, while the missing data rate and missing pattern are predetermined and treated as "random missing" events.

[0039] In step S3, different algorithms are used for point cloud data interpolation, including MICE, MissForest, inverse distance weighted KNN, texture-driven EKNN, and GAIN.

[0040] The specific steps for data interpolation using the inverse distance weighted KNN algorithm include: S3-11: For each missing point P m ( x m , y m Identify the set of candidate neighbor points in a two-dimensional raster matrix. N m The similarity between the missing point and each candidate neighbor point is calculated using Euclidean distance.

[0041] S3-12: Based on similarity ranking, select the K most similar neighbor points as the nearest neighbor set. P 1, P 2,…, P K}

[0042] S3-13: Elevation of the selected K neighbor points z k Perform a weighted average to obtain the missing points. P m Elevation estimate z m The weighting coefficients employ an inverse distance weighting strategy, with the weights defined as follows: (2) In the formula, d i missing points P m With the i Neighboring points P i The Euclidean distance between them δ To prevent constants with a denominator of zero, 10 is preferred. -6 .

[0043] The elevation values ​​of the missing points are obtained based on the above weights. z m That is, the estimated missing points: (3) S3-14: Incorporate the estimated missing points into the neighborhood search, repeat steps S3-12 and S3-13, and iterate the imputation until all missing values ​​are estimated.

[0044] S3-15: Perform local planar or bilinear smoothing on the interpolated point cloud to ensure elevation continuity.

[0045] The specific steps for data interpolation using the texture-driven EKNN algorithm include: S3-21: Add the observed valid points from the point cloud dataset { P i ( x i , y i , z i As a neighborhood search sample set, the missing points { P m} as the target to be estimated.

[0046] S3-22: Initialize the K value, set the initial value K=3, and allow the search range to be 3~10.

[0047] S3-23: Optimize parameter K using the discrete grid search method. The specific steps are as follows: S3-231: Establish a discrete candidate set K={3,4,5,6,7,8,9,10} within the search interval K∈[3,10]. For each candidate K value, use the KNN method to initially fill in the missing points and generate complete point cloud data.

[0048] S3-232: Calculate key road texture indicators based on the restored point cloud data, including mean texture depth (MTD), skewness (Ssk), and kurtosis (Sku), such as... Figure 3 The diagram shown illustrates the representation and calculation method of road surface texture index.

[0049] S3-233: Construct a texture-driven fitness function as shown in formula (4), compare the texture index of the recovered point cloud with the reference value (adjacent complete region or historical standard), and calculate the fitness value corresponding to the candidate K value.

[0050] (4) In the formula, J ( K ) is the fitness function, used to measure the consistency between the interpolation result and the texture features of the reference region under different K values; q rec ( KThe pavement texture indices (including mean texture depth MTD, skewness Ssk, and kurtosis Sku) are calculated after interpolation using candidate parameter K. q ref Reference texture index for adjacent complete regions or historical standard data; ε To prevent small constants with denominators of zero, we take 10. -6 ; ω q The weighting coefficients for each texture index satisfy the normalization condition: (5) Preferably, the weight is set to... ω MTD =0.4, ω Ssk =0.3, ω Sku =0.3.

[0051] S3-234: Select the K value with the smallest fitness from the candidate set as the optimal K value. opt .

[0052] S3-24: Using the optimal K value K opt The missing points are then interpolated to generate complete point cloud data.

[0053] S3-25: Perform local planar or bilinear smoothing on the interpolated point cloud to ensure elevation continuity.

[0054] Step S4 specifically includes: S41: The complete point cloud data after interpolation and repair is divided into blocks according to a fixed grid. For each local area, a continuous road surface elevation surface model is established by combining local quadratic surface fitting and triangular mesh (TIN) generation. Specifically, firstly, local data blocks are divided using a fixed-size sliding window (15×15 grid). Within each data block, the least squares method is used to fit a local quadratic surface to obtain a smooth elevation change expression. Then, using the center point of each fitted surface as a control node, a global triangular mesh is generated using the Delaunay triangulation algorithm to achieve geometric splicing and continuous transition between adjacent local surfaces.

[0055] S42: Based on the reconstructed 3D pavement model, calculate macroscopic and microscopic texture feature parameters for evaluating pavement function, including mean texture depth (MTD), skewness (Ssk), and kurtosis (Sku).

[0056] The root mean square error (RMSE) of pavement texture indices generated by interpolated data and real data is used to evaluate the pavement texture reconstruction effect of different interpolation algorithms.

[0057] like Figure 4 The figure shows a comparison of the reconstruction accuracy of different interpolation algorithms for the MTD index. It can be seen from the figure that the MTD error value increases with the increase of the missing rate. When the missing rate is less than 10%, the MICE, MissForest, and KNN algorithms show high accuracy in MTD reconstruction; however, the KNN algorithm is superior. When the missing rate exceeds 30%, the EKNN algorithm shows higher MTD reconstruction capability. GAIN performs poorly in MTD reconstruction at all missing rate levels.

[0058] like Figure 5 The figure shows a comparison of the reconstruction accuracy of the Ssk index by different imputation algorithms. As can be seen from the figure, the Ssk error value gradually increases with the increase of the missing rate. When the missing rate is less than 10%, the MICE, MissForest, and KNN algorithms all show high accuracy in reconstructing the Ssk, with the KNN algorithm being superior. When the missing rate exceeds 30%, the EKNN algorithm shows higher Ssk reconstruction ability. GAIN performs poorly in Ssk reconstruction at all missing rate levels.

[0059] like Figure 6 The figure shows a comparison of the reconstruction accuracy of different interpolation algorithms for the SKU index. As can be seen from the figure, the SKU error value gradually increases with the increase of the missing rate. When the missing rate is less than 10%, the MICE, MissForest, and KNN algorithms all show high accuracy in reconstructing the SKU. When the missing rate exceeds 30%, GAIN shows higher SKU reconstruction capability, while the EKNN algorithm is slightly inferior.

[0060] Table 1 compares the computational efficiency of different interpolation algorithms. As can be seen from the table, KNN and texture-driven EKNN algorithms significantly outperform other methods in terms of computational efficiency.

[0061]

[0062] When the data missing rate is less than 10%, both MICE and MissForest show good accuracy in reconstructing road surface texture indicators, but their computational efficiency is significantly lower. Therefore, when the missing rate is less than 10%, the inverse distance KNN algorithm used in this invention performs best. When the missing rate is greater than 10%, the texture-driven EKNN algorithm proposed in this invention shows better texture reconstruction accuracy and higher computational efficiency, thus performing best. In summary, the method proposed in this invention achieves a good balance between reconstruction effect and efficiency.

[0063] S43: To further improve the realism of texture restoration, texture optimization and consistency correction are performed on the reconstructed surface. First, Gaussian smoothing is used to denoise the surface and eliminate isolated elevation outliers caused by the interpolation algorithm. Second, macroscopic and microscopic texture parameters are compared with the statistical characteristics of adjacent complete regions. If the deviation exceeds a preset threshold, local refitting is performed to ensure the spatial continuity and consistency of texture indicators.

[0064] S44: Outputs the optimized road surface data as a standard 3D model format for subsequent structural analysis and performance evaluation.

[0065] The reconstructed profile elevation distribution of asphalt concrete pavement with a 10% rectangular data missing area is as follows: Figure 7 , Figure 8 As shown, Figure 7 and Figure 8 respectively along x shaft and y The distribution of profile elevation along the axial direction, along x shaft and y The root mean square errors of the profile elevation in the axial direction are 0.28 mm and 0.50 mm, respectively. Therefore, the method proposed in this invention can better restore the texture features of the road surface.

[0066] Example 2 like Figure 9 As shown, based on the same inventive concept as the above embodiments, the present invention also provides a road surface texture reconstruction system based on vehicle-mounted laser point cloud missing data interpolation, comprising: The acquisition unit is used to acquire road surface point cloud data and perform preprocessing to obtain the effective analysis area and effective road surface point cloud data. A calculation unit is used to determine the missing rate based on the effective analysis area and effective road surface point cloud data; The interpolation unit is used to determine the interpolation algorithm based on the missing rate and to repair the point cloud data to obtain complete point cloud data. The reconstruction unit is used to sequentially construct a 3D road surface model based on complete point cloud data, calculate multi-scale texture features, perform texture optimization and consistency correction, and output road surface data.

[0067] Example 3 like Figure 10 As shown, the present invention also provides an electronic device 100 for implementing a road surface texture reconstruction method based on missing data interpolation of vehicle-mounted laser point cloud; The electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on at least one processor 102, and at least one communication bus 104.

[0068] The memory 101 can be used to store the computer program 103. The processor 102 implements the steps of the road texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101.

[0069] The memory 101 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0070] At least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 102 may be a microprocessor or any conventional processor. Processor 102 is the control center of electronic device 100, connecting various parts of electronic device 100 via various interfaces and lines.

[0071] The memory 101 in the electronic device 100 stores multiple instructions to implement a road texture reconstruction method based on missing data interpolation of vehicle-mounted laser point cloud, and the processor 102 can execute multiple instructions to achieve the following: Acquire road surface point cloud data and perform preprocessing to obtain the effective analysis area and effective road surface point cloud data; The missing rate is determined based on the effective analysis area and effective road surface point cloud data; The imputation algorithm is determined based on the missing rate, and the point cloud data is repaired to obtain complete point cloud data; Based on complete point cloud data, a 3D road surface model is constructed sequentially, multi-scale texture features are calculated, texture optimization and consistency correction are performed, and road surface data is output. Example 4 If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).

[0072] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0073] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0074] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0075] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0076] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A road surface texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation, characterized in that, The method comprises the following steps: Obtain road point cloud data and preprocess it to obtain an effective analysis area and effective road point cloud data; Determine the missing rate based on the effective analysis area and the effective road point cloud data; Determine the interpolation algorithm according to the missing rate, and repair the point cloud data to obtain complete point cloud data; Based on the complete point cloud data, a three-dimensional road surface model is constructed, multi-scale texture features are calculated, and texture optimization and consistency correction are performed, and finally road surface curved surface data is output.

2. The road texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation according to claim 1, characterized in that, The preprocessing comprises: Extract a central rectangular area from the collected road point cloud data as the effective analysis area, and perform noise reduction and smoothing processing on the point cloud data in the area to eliminate abnormal points, thereby obtaining effective road point cloud data.

3. The road texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation according to claim 1, characterized in that, The missing rate is determined based on the effective analysis area and the effective road point cloud data, which comprises: Divide the effective analysis area into a grid to establish a regular two-dimensional coordinate grid matrix, and map each effective point to the corresponding two-dimensional coordinate grid matrix according to the spatial coordinates of the point cloud; Identify all missing points in the two-dimensional coordinate grid matrix and count the number of missing points; Calculate the missing rate based on the number of missing points and the effective road point cloud data.

4. The road texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation according to claim 1, characterized in that, The interpolation algorithm is determined according to the missing rate, and the point cloud data is repaired to obtain complete point cloud data, which comprises: When the missing rate is less than or equal to 10%, the point cloud data is repaired based on the inverse distance weighted KNN algorithm to obtain complete point cloud data; When the missing rate is greater than 10%, the point cloud data is repaired based on the texture-driven EKNN algorithm to obtain complete point cloud data.

5. The method of claim 4, wherein, The point cloud data is repaired based on the inverse distance weighted KNN algorithm, which comprises: For each missing point, identify its candidate neighbor point set in the two-dimensional grid matrix, and calculate the similarity between the missing point and each candidate neighbor point; According to the similarity ranking, select K most similar neighbor points as the nearest neighbor set; Perform a weighted average on the elevations of the selected K neighbor points to calculate the missing point elevation value, thereby obtaining the estimated missing point; Include the estimated missing point in the neighborhood search, and repeat the above steps to iteratively interpolate until all missing points are estimated; Perform local plane or bilinear smoothing on the interpolated point cloud.

6. The method of claim 4, wherein, The point cloud data is repaired based on the texture-driven EKNN algorithm, which comprises: The observed effective points in the point cloud data set are used as the neighborhood search sample set, and the missing points are used as the target to be estimated; Initialize the K value, set the initial value K=3, and set the search range to 3-10; Optimize the parameter K using the discrete grid search method to obtain the optimal K value; Use the optimal K value to perform final interpolation on the missing points to generate complete point cloud data; Perform local plane or bilinear smoothing on the complete point cloud data.

7. The method of claim 1, wherein, Based on the complete point cloud data, a three-dimensional road surface model is constructed, multi-scale texture features are calculated, and texture optimization and consistency correction are performed, and finally road surface curved surface data is output, which comprises: Based on the complete point cloud data, a three-dimensional road surface model is constructed, multi-scale texture features are calculated, and texture optimization and consistency correction are performed, and finally road surface curved surface data is output, which comprises: Based on the complete point cloud data, perform block processing according to a fixed grid to obtain a plurality of local areas, and for each local area, a continuous three-dimensional road surface model is established by combining local quadratic surface fitting and triangular mesh generation; Based on the three-dimensional road surface model, calculate macro and micro texture feature parameters for evaluating road surface function; Based on macro and micro texture feature parameters, the three-dimensional road surface model is optimized and corrected in consistency to obtain the optimized road surface curved surface data. The optimized road surface curved surface data is output in a three-dimensional model format. 8.A road surface texture reconstruction system based on missing data interpolation of vehicle-mounted laser point cloud, characterized in that, The method comprises the following steps: An acquisition unit is configured to acquire road point cloud data and perform preprocessing to obtain an effective analysis region and effective road point cloud data; A calculation unit is configured to determine a missing rate based on the effective analysis region and the effective road point cloud data; An interpolation unit is configured to determine an interpolation algorithm according to the missing rate and perform point cloud data repair to obtain complete point cloud data; A reconstruction unit is configured to sequentially construct a three-dimensional road surface model, calculate multi-scale texture features, and perform texture optimization and consistency correction based on the complete point cloud data, and finally output road surface curved surface data.

9. An electronic device, comprising: The processor is configured to execute a computer program stored in the memory to implement the road surface texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores at least one instruction, and the at least one instruction is executed by the processor to implement the road surface texture reconstruction method based on vehicle-mounted laser point cloud missing data interpolation according to any one of claims 1 to 7.

Images