Point cloud repairing method and system and storage medium

Abstract

The invention discloses a point cloud repairing method. The method comprises the following steps of S1, starting, S2, data acquisition, S3, point cloud preprocessing, S4, point cloud projection, S5, B spline fitting, S6, point cloud data table generation, S7, coordinate searching, S8, interpolation processing and S9, output processing. And S10, ending. According to the point cloud repairing system based on the combination of the B spline and the bilinear interpolation, the repairing precision and the processing efficiency of the point cloud data can be effectively improved, and the high requirement of modern engineering application for three-dimensional data processing is met.

Description

Technical Field

[0001] The present invention relates to point cloud processing and data repair technology, and in particular to the field of a method based on combining B-spline and bilinear interpolation. Background Art

[0002] In the field of point cloud processing, the existing technology mainly includes data repair technology based on traditional interpolation methods and curve fitting algorithms. For example, the public document with announcement number CN114612632A, publication date June 10, 2022, and patent name "A sorting and interpolation processing method based on three-dimensional laser point cloud data" discloses a sorting and interpolation processing method based on three-dimensional laser point cloud data, including: S1, data acquisition: collect point cloud data with a three-dimensional laser scanner; S2, cropping: crop and delete the collected non-target points; S3, repair: perform preliminary repair on the three-dimensional image to fill holes; S4, cutting: cut the point cloud data into standard rectangles for ease of calculation; S5, labeling: divide the grid according to the coordinates of the rectangle vertices and sort the points by labeling; S6, interpolation: first use the coordinates and connection relationship of the midpoint of the spatial triangle surface for interpolation, and then use the information of the eight points around each point for interpolation.

[0003] These methods usually use linear interpolation or polynomial fitting to fill in missing data and are widely used in computer vision, 3D modeling, and reverse engineering. However, with the advancement of point cloud data acquisition technology, the amount and complexity of acquired data have increased significantly, and the limitations of traditional methods have gradually been exposed.

[0004] Current point cloud repair technologies mostly rely on simple interpolation algorithms, which can quickly generate estimates of missing parts when processing point cloud data. However, they are usually unable to effectively capture subtle changes and complex structures in the data, especially when processing data with high curvature or complex surfaces. It is difficult to provide satisfactory repair results. In addition, existing technologies often require a lot of manual intervention and post-processing, which increases the workload and time cost. Therefore, there is an urgent need to develop new methods to meet the needs of modern point cloud processing and improve the automation and intelligence level of data repair.

[0005] In general, the disadvantages of the prior art include:

[0006] Insufficient smoothness: Traditional interpolation methods often cannot effectively handle complex surfaces, resulting in obvious jagged or rough appearance in the repaired data.

[0007] Locality problem: Many interpolation algorithms only consider the information of neighboring points and lack a global perspective, which may lead to inaccurate restoration results when large-scale data is missing.

[0008] Low computational efficiency: For large-scale point cloud data, traditional algorithms are often computationally inefficient and cannot meet the needs of real-time processing.

[0009] Sensitivity to noise: Traditional methods are sensitive to noise in the data, which may lead to erroneous interpolation results and further affect data quality.

[0010] Lack of adaptability: Existing technologies are usually based on fixed model assumptions and are difficult to adapt to the characteristics of different data sets, limiting their wide application.

[0011] Long processing time: When faced with complex scenes or high-density point clouds, traditional algorithms take a long time to process and are not suitable for fast application scenarios. Summary of the invention

[0012] The technical problem to be solved by the present invention is to realize a point cloud repair method which combines B-spline and bilinear interpolation to improve the accuracy and efficiency of point cloud repair.

[0013] In order to achieve the above object, the technical solution adopted by the present invention is: a point cloud repair method, comprising the following steps:

[0014] S1. Start: System initialization, ready for point cloud data processing;

[0015] S2, data collection: collect point cloud data to obtain three-dimensional data;

[0016] S3, point cloud preprocessing: performing noise filtering, point cloud density adjustment and format conversion;

[0017] S4, point cloud projection: project the 3D point cloud onto a 2D plane and extract the X and Y coordinates;

[0018] S5, B-spline fitting: Generate smooth curves using the B-spline algorithm and perform fitting error evaluation;

[0019] S6, point cloud data table generation: extract valid point cloud data and record the corresponding Z coordinates;

[0020] S7, coordinate search: search for the corresponding Z coordinate value based on the known X and Y coordinates;

[0021] S8, interpolation processing: calculating the missing Z-sit by bilinear interpolation algorithm;

[0022] S9, output processing: export the repaired point cloud data and perform visual display;

[0023] Save the processed data in a usable format and provide an intuitive visualization interface;

[0024] S10, end.

[0025] In S2, a line laser sensor is used to measure the distance of the surface of the object, and high-density point cloud data is generated for point cloud data collection to obtain high-precision three-dimensional data.

[0026] In S3, when noise is filtered, outliers and interfering data are removed by an algorithm, and when point cloud density is adjusted, data distribution is optimized according to demand.

[0027] In S4, a smooth curve is generated using the B-spline algorithm:

[0028]

[0029] in:

[0030] C(t) is the point of the B-spline curve under parameter t;

[0031] n is the number of control points minus 1, which means, n = m-1, where m is the total number of control points;

[0032] P i is the i-th control point;

[0033] N i，p (t) is the i-th B-spline basis function with order p.

[0034] In S5, a continuous curve is generated by a fitting algorithm to optimize the shape of the point cloud data and reduce unnecessary fluctuations.

[0035] In S8, the values ​​of unknown points are inferred from the values ​​of known points to fill in the data gaps;

[0036] The specific steps are:

[0037] Step 1: Select four adjacent known coordinates V 00 , V 01 , V 10 , V 11 , each position stores the current Z;

[0038] Step 2: Use the bilinear interpolation formula to calculate the compensation value of the target position based on the error value of the known point. The model is as follows:

[0039]

[0040] Where v(x, y) represents the height of the point cloud, and X and Y are the side lengths in the x and y directions, respectively.

[0041] In S10, the system completes all processing steps and is ready to close or restart a new data processing task.

[0042] A point cloud repair system includes a three-axis platform, a camera for shooting downwards is provided above the three-axis platform, the camera is used for collecting three-dimensional point cloud data, the camera is connected to and outputs a sensing signal to a computer, and the computer is used to execute all software and algorithm modules, and perform data processing, storage and user interaction.

[0043] The computer comprises:

[0044] Data acquisition module: used to receive data from the line laser camera;

[0045] Data output module: It is used to save the processed point cloud data into a specified format and interact with the user;

[0046] Communication module: responsible for data transmission between hardware devices and computers;

[0047] Point cloud filtering module: used for noise filtering, point cloud density adjustment and format conversion;

[0048] Projection module: used to project 3D point cloud data onto a 2D plane and extract X and Y coordinates;

[0049] B-spline fitting module: used to generate a smooth curve using the B-spline algorithm and pass the generated fitting results to the interpolation module and the data export module.

[0050] A storage medium is a computer-readable storage medium for storing software program codes, wherein the software program codes are used to execute the point cloud repair method.

[0051] The advantages of the present invention are:

[0052] Improve the repair accuracy. By combining B-spline and bilinear interpolation, the system can accurately repair the missing point cloud data and ensure that the generated point cloud has higher spatial accuracy and continuity;

[0053] Enhanced data integrity: This system can effectively fill in missing Z coordinates to ensure the integrity of point cloud data, thereby improving the reliability of subsequent analysis and visualization;

[0054] Local control capability: B-spline has good local control characteristics. Moving a control point only affects the repair effect of its adjacent area, allowing users to flexibly adjust and optimize data.

[0055] Smoothness and visualization effect: The curve generated by the B-spline algorithm has high smoothness, which can effectively reduce the fluctuation in the point cloud data and enhance the visualization effect, making the final display result more natural and realistic;

[0056] Real-time and high efficiency: The design of this system takes into account the real-time performance of the algorithm, so that when processing large-scale point cloud data, it can still maintain a high computing speed and meet the needs of modern engineering applications;

[0057] Wide applicability: The point cloud repair system is suitable for a variety of application scenarios, such as reverse engineering, 3D modeling, and industrial inspection, and can meet the needs of data repair in different fields;

[0058] The simplified workflow combines the advantages of the two algorithms. This system can simplify the data processing process, reduce manual intervention, reduce operational complexity, and improve user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The following is a brief description of the contents expressed in each figure in the specification of the present invention and the marks in the figure:

[0060] Figure 1 A framework diagram for the point cloud repair system of the present invention;

[0061] Figure 2 It is a schematic diagram of the structure of the point cloud repair system of the present invention;

[0062] Figure 3 This is a workflow diagram of the point cloud repair system of the present invention;

[0063] Figure 4 This is a point cloud filtering effect diagram of the present invention;

[0064] Figure 5 The projection and spline fitting diagram of the point cloud of the present invention;

[0065] Figure 6 This is a schematic diagram of the point cloud repair interpolation principle of the present invention;

[0066] Figure 7 This is the mesh effect diagram after the point cloud is repaired in the present invention;

[0067] The marks in the above figures are: 1, X-axis; 2, Y-axis; 3, Z-axis; 4, camera. DETAILED DESCRIPTION

[0068] The following is a further detailed description of the specific implementation methods of the present invention, such as the shape, structure, relative position and connection relationship between the various components involved, the function and working principle of each part, the manufacturing process and operation method, etc., through the description of the embodiments with reference to the accompanying drawings, so as to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.

[0069] The shortcomings of existing point cloud repair methods mainly stem from the simple assumptions of traditional methods on data features, which fail to fully utilize the geometric characteristics and global information of point cloud data. The lack of flexible mathematical models and effective algorithms makes it difficult to adapt to complex 3D shape changes and data incompleteness. Therefore, a new method combining B-spline and bilinear interpolation is urgently needed to improve the accuracy and efficiency of point cloud repair.

[0070] The point cloud repair system based on the combination of B-spline and bilinear interpolation projects the current frame point cloud into two-dimensional space, performs B-spline fitting, uses the x and y coordinates of the fitting curve, and generates a point cloud data table with the filtered valid point cloud. The coordinate of Z is found according to (x, y). If (x, y) does not exist, bilinear interpolation is used to calculate the current Z, which effectively solves the problems of insufficient accuracy, incomplete data and low processing efficiency in the existing point cloud data repair process.

[0071] like Figure 1 As shown, the system includes a hardware module, a software module and an algorithm module.

[0072] 1. The hardware module is used for continuous acquisition and storage of line laser point clouds;

[0073] Line laser camera 4: responsible for high-precision 3D point cloud data acquisition, obtains depth information of the object surface through laser ranging technology, provides raw data for the data acquisition module, and directly affects the subsequent data processing and repair effect. The device is installed at the end of the Z axis to ensure scanning at different heights to obtain comprehensive data.

[0074] Three-axis platform: used to provide a stable motion platform, so that the line laser camera 4 can scan at different angles and positions, ensure the comprehensiveness of the data, support the movement of the line laser camera 4, and affect the range and accuracy of data collection.

[0075] Computer: used to execute all software and algorithm modules, perform data processing, storage and user interaction. As the core of the system, it connects all hardware modules and software modules and is responsible for data transmission and processing.

[0076] 2. Software modules, which are used to manage data collection, processing and output, and support the interaction between users and the system, including:

[0077] Data acquisition module: It is used to receive data from the line laser camera 4, perform preliminary processing and storage, ensure the real-time and accuracy of the data, and provide reliable raw data for subsequent processing.

[0078] Data output module: It is used to save the processed point cloud data in a specified format for subsequent use; the visual display obtains the processing results from the data output module and interacts with the user to provide an intuitive visual interface.

[0079] Communication module: It is responsible for data transmission between hardware devices and computers, ensuring the real-time and accuracy of data and optimizing the overall performance of the system.

[0080] 3. The algorithm module is used to perform preprocessing, fitting and interpolation calculations of point cloud data to optimize data quality and fill in missing information, including:

[0081] Point cloud filtering module: It is used for noise filtering, point cloud density adjustment and format conversion to improve data quality and prepare high-quality data for subsequent projection and fitting modules.

[0082] Projection module: It is used to project the 3D point cloud data onto a 2D plane, extract the X and Y coordinates for further processing, receive data from the point cloud filtering module, and pass the results to the B-spline fitting module.

[0083] B-spline fitting module: It is used to generate a smooth curve using the B-spline algorithm to improve the continuity of the point cloud data, receive data from the projection module, and pass the generated fitting results to the interpolation module and the data export module.

[0084] Interpolation module:

[0085] Coordinate search: It is used to find the corresponding Z coordinate value based on the known X and Y coordinates to ensure data integrity.

[0086] Bilinear interpolation: It is used to calculate the corresponding Z value for the missing (X, Y) coordinates using the bilinear interpolation algorithm to fill in the data gaps and ensure the continuity and integrity of the point cloud data.

[0087] The workflow of the point cloud repair system based on the combination of B-spline and bilinear interpolation includes the following steps:

[0088] S1. Start: System initialization, ready for point cloud data processing.

[0089] S2. Data collection: Figure 2 As shown, the three-axis machine is moved and a line laser sensor is used to collect point cloud data to obtain high-precision three-dimensional data. This module is responsible for ensuring the accuracy of the data. Through laser emission and reception technology, the distance of the object surface is accurately measured to generate high-density point cloud data.

[0090] S3, point cloud preprocessing: perform noise filtering, point cloud density adjustment and format conversion to improve data quality. The noise filtering step removes outliers and interference data through algorithms to ensure the effectiveness of subsequent processing; point cloud density adjustment optimizes the distribution of data according to needs to adapt to different analysis situations, such as Figure 4 shown.

[0091] S4, point cloud projection: Project the 3D point cloud onto a 2D plane and extract the X and Y coordinates for subsequent processing. This step converts the 3D coordinates into a 2D representation through calculation, facilitating subsequent analysis and visualization while retaining the necessary spatial information, such as Figure 5 shown.

[0092] S5, B-spline fitting: Generate smooth curves using the B-spline algorithm and evaluate fitting errors. This module generates continuous curves through fitting algorithms, optimizes the shape of point cloud data, and reduces unnecessary fluctuations to improve visual effects and data usability.

[0093] The specific method is to use the B-spline algorithm to generate a smooth curve:

[0094]

[0095] in:

[0096] C(t) is the point of the B-spline curve under parameter t;

[0097] n is the number of control points minus 1, which means, n = m-1, where m is the total number of control points;

[0098] P i is the i-th control point;

[0099] N i，p (t) is the i-th B-spline basis function with order p.

[0100] This module generates continuous curves through fitting algorithms, optimizes the shape of point cloud data, and reduces unnecessary fluctuations to improve visual effects and data usability, such as Figure 5 shown.

[0101] S6. Point cloud data table generation: Extract valid point cloud data and record the corresponding Z coordinates. This step ensures that the extracted data is complete and accurate, provides the necessary basic data for subsequent interpolation calculations, and ensures that the generated point cloud data table is complete.

[0102] S7, Coordinate Search: Find the corresponding Z coordinate value based on the known X and Y coordinates. This module quickly locates the missing Z value through the search algorithm, ensuring the integrity and continuity of the data and providing accurate input for the interpolation process.

[0103] S8, interpolation module: Calculate the missing Z coordinates through the bilinear interpolation algorithm. The algorithm infers the values ​​of unknown points through the values ​​of known points, fills in the data gaps, and enhances the quality and integrity of the final point cloud. The error values ​​in the x and y directions are used as reference values ​​to calculate the z value. The specific steps are as follows:

[0104] The specific steps are:

[0105] Step 1: Select four adjacent known coordinates such as Figure 6 Shown v 00 , v 01 , v 10 , v 11 , each position stores the current Z;

[0106] Step 2: Use the bilinear interpolation formula to calculate the compensation value of the target position based on the error value of the known point. The model is as follows:

[0107]

[0108] Where v(x, y) represents the height of the point cloud, and X and Y are the side lengths in the x and y directions, respectively.

[0109] S9, Output module: Export the repaired point cloud data and display it visually for easy analysis and use by users. This module saves the processed data in a usable format and provides an intuitive visualization interface to help users understand and analyze the results.

[0110] S10, End: The system completes all processing steps and is ready to close or restart a new data processing task.

[0111] Through the above steps, the system realizes the efficient collection, processing and repair of 3D point cloud data, ensuring the integrity and accuracy of the data, while providing intuitive visual display, meeting the high requirements of modern engineering applications for data quality and processing efficiency.

[0112] The present invention combines B-spline and bilinear interpolation, uniquely combining the B-spline algorithm with bilinear interpolation to form a new point cloud patching method that can efficiently and accurately fill in missing point cloud data. Multi-module system architecture The system design includes hardware modules, software modules and algorithm modules to form a complete solution to ensure the efficiency and accuracy of data acquisition, processing and output.

[0113] Local control characteristics: Using the local control characteristics of B-spline, users can flexibly adjust the control points to optimize the point cloud repair effect and improve the usability and adaptability of the system. Efficient data processing flow: The system optimizes the algorithm in multiple steps such as data preprocessing, projection, fitting and interpolation, improves the processing speed and real-time performance, and adapts to the needs of large-scale point cloud data. The wide range of applications: It can be applied to different application fields, such as reverse engineering, 3D modeling, industrial inspection, etc., with versatility and flexibility.

[0114] The present invention is described above by way of example in conjunction with the accompanying drawings. It is obvious that the specific implementation of the present invention is not limited to the above-mentioned method. As long as various non-substantial improvements are made using the method concept and technical solution of the present invention, or the concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A point cloud repair method, characterized in that: The following steps are involved: S1. Start: System initialization, ready for point cloud data processing; S2, data collection: collect point cloud data to obtain three-dimensional data; S3, point cloud preprocessing: performing noise filtering, point cloud density adjustment and format conversion; S4, point cloud projection: project the 3D point cloud onto a 2D plane and extract the X and Y coordinates; S5, B-spline fitting: Generate smooth curves using the B-spline algorithm and perform fitting error evaluation; S6, point cloud data table generation: extract valid point cloud data and record the corresponding Z coordinates; S7, coordinate search: search for the corresponding Z coordinate value based on the known X and Y coordinates; S8, interpolation processing: calculating the missing Z-sit by bilinear interpolation algorithm; S9, output processing: export the repaired point cloud data and perform visual display; Save the processed data in a usable format and provide an intuitive visualization interface; S10, end.

2. The point cloud repair method according to claim 1, characterized in that: In S2, a line laser sensor is used to measure the distance of the surface of the object, and high-density point cloud data is generated for point cloud data collection to obtain high-precision three-dimensional data.

3. The point cloud repair method according to claim 1, characterized in that: In S3, when noise is filtered, outliers and interfering data are removed by an algorithm, and when point cloud density is adjusted, data distribution is optimized according to demand.

4. The point cloud repair method according to claim 1, characterized in that: In S4, a smooth curve is generated using the B-spline algorithm: in: C(t) is the point of the B-spline curve under parameter t; n is the number of control points minus 1, which means, n = m-1, where m is the total number of control points; P i is the i-th control point; N i，p (t) is the i-th B-spline basis function with order p.

5. The point cloud repair method according to claim 1, characterized in that: In S5, a continuous curve is generated by a fitting algorithm to optimize the shape of the point cloud data and reduce unnecessary fluctuations.

6. The point cloud repair method according to claim 1, characterized in that: In S8, the values ​​of unknown points are inferred from the values ​​of known points to fill in the data gaps; The specific steps are: Step 1: Select four nearby known coordinates v 00 , v 01 , v 10 , v 11 , each position stores the current Z; Step 2: Use the bilinear interpolation formula to calculate the compensation value of the target position based on the error value of the known point. The model is as follows: Where v(x, y) represents the height of the point cloud, and X and Y are the side lengths in the x and y directions, respectively.

7. The point cloud repair method according to claim 1, characterized in that: In S10, the system completes all processing steps and is ready to close or restart a new data processing task.

8. A point cloud repair system, characterized in that: The system includes a three-axis platform, on which a downward-shooting camera is provided. The camera is used for collecting three-dimensional point cloud data. The camera is connected to and outputs sensing signals to a computer. The computer is used to execute all software and algorithm modules, and perform data processing, storage and user interaction.

9. The point cloud repair system according to claim 8, characterized in that: The computer comprises: Data acquisition module: used to receive data from the line laser camera; Data output module: It is used to save the processed point cloud data into a specified format and interact with the user; Communication module: responsible for data transmission between hardware devices and computers; Point cloud filtering module: used for noise filtering, point cloud density adjustment and format conversion; Projection module: used to project 3D point cloud data onto a 2D plane and extract X and Y coordinates; B-spline fitting module: used to generate a smooth curve using the B-spline algorithm and pass the generated fitting results to the interpolation module and the data export module.

10. A storage medium, the storage medium being a computer-readable storage medium for storing software program codes, characterized in that: The software program code is used to execute the point cloud repair method as described in any one of claims 1-7.

Images