Browser-based laser point cloud data online resolving and processing method

Abstract

The invention discloses a browser-based laser point cloud data online resolving and processing method. The method comprises the following steps: data loading and displaying; data preprocessing and filtering; feature data extraction and effective analysis; cloud computing support is completed; interactive analysis and visualization processing are completed; according to the laser point cloud data online resolving and processing method based on the browser, a large amount of laser point cloud data can be effectively recorded in a modern Web browser, display and interaction are completed in real time, effective point cloud data analysis and visual experience are provided for a user, and meanwhile, the architecture and process design involved in the method is easy to implement and convenient to use. Large-scale point cloud data can be efficiently processed and analyzed, task management and scheduling are achieved on a browser interface, the real-time interaction requirement of a user for multi-dimensional data processing is met, and data export and sharing functions are supported.

Description

Technical Field

[0001] The invention relates to the technical field of laser point cloud processing, and in particular to an online solution and processing method for laser point cloud data based on a browser. Background Art

[0002] LiDAR point cloud data is a type of spatial data obtained through laser scanning, which has the characteristics of high accuracy and rich information. LiDAR point cloud data processing is the process of processing, analyzing and mining these data. LiDAR point cloud data processing plays an important role in the fields of 3D modeling, unmanned driving, robotics, etc. Taking unmanned vehicles as an example, LiDAR point cloud data processing can be applied to the environmental perception and navigation control of unmanned vehicles. By obtaining point cloud data of the surrounding environment through LiDAR scanning, data preprocessing and segmentation, the environment around the vehicle can be identified and classified. Then 3D modeling is performed to build a 3D environment model around the vehicle, and further intelligent driving tasks such as path planning and obstacle avoidance control are performed. Finally, the processed point cloud data is fused and exported to realize the application of unmanned vehicles in actual scenarios.

[0003] The existing technology has the following defects or problems: the existing laser point cloud data processing methods are mostly based on application clients, the installation and upgrade procedures are cumbersome, and the equipment replacement is troublesome; at the same time, the interactive effects are cross-cutting, which cannot meet the user's needs for multi-dimensional data processing and real-time interaction, and the user experience is poor. Summary of the invention

[0004] The purpose of the present invention is to provide a browser-based online solution and processing method for laser point cloud data in view of the shortcomings of the prior art, so as to solve the problems raised in the background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a browser-based online solution and processing method for laser point cloud data, comprising the following steps:

[0006] Step 1: Data loading and display:

[0007] Use WebGL technology to load a large amount of laser point cloud data into the browser, and display and interact in real time through 3D scene rendering technology;

[0008] Step 2: Data preprocessing and filtering:

[0009] Implement preprocessing of laser point cloud data in the browser, such as denoising, filtering and registration operations;

[0010] Step 3: Feature data extraction and effective analysis:

[0011] Browser-based algorithms enable feature data extraction from point clouds, such as ground extraction, object recognition and classification;

[0012] Step 4: Real-time solution and model fitting:

[0013] Use the computing power in the browser to perform real-time settlement and model fitting of point cloud data;

[0014] Step 5: Complete cloud computing support:

[0015] Combined with cloud computing resources, it can realize efficient processing and analysis of large-scale point cloud data, and manage and schedule computing tasks through the browser interface;

[0016] Step 6: Complete interactive analysis and visualization:

[0017] It provides a user-friendly interactive interface, allowing users to perform multi-dimensional analysis and real-time visualization of point cloud data in the browser, and supports data export and sharing.

[0018] As a preferred technical solution of the present invention, the specific operation steps of data loading and display are as follows:

[0019] 1) Laser point cloud data is usually stored in LAS (Lidar Data Exchange Format) or PTS (Point Cloud Data) format. Before loading into WebGL, the following preprocessing may be required:

[0020] Data format conversion: Convert the original point cloud data format into a format that can be processed by WebGL, usually using JSON and binary data;

[0021] Data compression: compress data to reduce transmission and preloading time;

[0022] Coordinate system conversion: The coordinate system of the point cloud data can be converted into a coordinate system more suitable for WebGL rendering as needed;

[0023] 2) Efficiency in recording large amounts of point cloud data in WebGL is key, and there are usually several approaches:

[0024] Load directly into memory: For small-scale point cloud data, the data can be directly stored in JavaScript variables;

[0025] Block loading: For large-scale data, the block loading method is adopted to dynamically load and unload data blocks according to the field of view;

[0026] Using buffer objects: Buffer objects in WebGL manage and process large amounts of data at a constant and efficient speed. Through buffer mapping, point cloud data can be sent to the GPU for parallel processing and rendering.

[0027] 3) Since WebGL is a web graphics library based on OpenGL ES, it allows the use of JavaScript in web browsers for hardware-accelerated 3D graphics rendering. In point cloud data rendering, the following steps can be used to achieve it:

[0028] Vertex shader and fragment shader: The vertex shader is used to process the position and color information of the point, and the fragment shader is used to render the final color of each point;

[0029] Buffer Objects: Use buffers to manage vertex data and attribute data for efficient processing and rendering on the GPU;

[0030] Shader Programs: Write and optimize vertex shader and fragment shader programs to maximize rendering performance and quality;

[0031] 4) To complete real-time display and interaction, you can use the following methods:

[0032] View transformation and camera control: Enable users to rotate, zoom, and translate the scene, so that the operation can change the viewing angle;

[0033] Point selection and performance: realize the point selection and selection function of point cloud data, and detect the point cloud clicked by the user through ray picking;

[0034] Dynamic update: dynamically update the rendered point cloud data based on user interaction and data updates;

[0035] 5) When loading and rendering point cloud data, performance and security issues also need to be considered:

[0036] Data flow and segmented loading: Avoid recording too much data at one time. Use appropriate data flow and segmented loading strategies to improve rendering efficiency and user experience.

[0037] Ensure data is transmitted via HTTPS to prevent sensitive data from being leaked during transmission.

[0038] As a preferred technical solution of the present invention, the denoising and filtering process is specifically as follows:

[0039] 1) First, ensure that the laser point cloud data can be effectively loaded and parsed in WebGL;

[0040] 2) Select an effective denoising and filtering algorithm, which can be completed by using a statistical filter, a voxel grid filter, or a Gaussian filter;

[0041] 3) To implement the above algorithm in a browser environment, an efficient JavaScript code is usually required, which is completed by using a statistical filter;

[0042] 4) During the implementation process, debugging and optimization are required to ensure that the algorithm can be processed in the browser within a reasonable time;

[0043] The registration process is specifically as follows:

[0044] 1) It is necessary to ensure that two sets of point cloud data can be loaded and parsed, used as source point cloud and target point cloud respectively, and the data can be obtained through file loading or network request;

[0045] 2) Use the iterative closest point algorithm to find the optimal rotation and translation transformation so that the two sets of point clouds overlap as much as possible;

[0046] 3) Then use the appropriate WebGL library to render the processed point cloud data into the browser so that users can view the alignment results. In the above process, debugging and performance optimization are required.

[0047] As a preferred technical solution of the present invention, the ground extraction of point cloud data is specifically as follows:

[0048] 1) Complete the loading of point cloud data by loading files;

[0049] 2) Random Sampling Consensus (RANSAC) is used to complete ground extraction;

[0050] 3) After extracting ground and non-ground points, further processing of this type of data is completed, and the processed point cloud data is rendered to the browser using WebGL;

[0051] The ground extraction of point cloud data is specifically as follows:

[0052] 1) Complete the loading of point cloud data by loading files;

[0053] 2) Select the object segmentation method based on region production and implement object segmentation in the browser;

[0054] 3) After object recognition is completed, the identified object point cloud is processed and visualized;

[0055] The point cloud data classification process is specifically as follows:

[0056] 1) Complete the loading of point cloud data by loading files;

[0057] 2) Select the point cloud classification method based on support vector machine (SVM) to classify the point cloud data in the browser;

[0058] 3) After completing the point cloud classification, the classification results can be processed and visualized.

[0059] As a preferred technical solution of the present invention, the real-time settlement and model fitting steps are specifically as follows:

[0060] 1) Complete the loading of the pre-processed point cloud data in the above steps;

[0061] 2) For point cloud data, it is necessary to extract features and calculate descriptors. Feature extraction can be based on the geometric properties, normal vectors, and curvature of the point cloud, while the descriptor is a mathematical description of the local or global structure;

[0062] 3) Select the appropriate model according to the application scenario;

[0063] 4) Using a numerical optimization algorithm to fit the model to the point cloud data, the numerical optimization algorithm may be one of the least squares method and the gradient descent method;

[0064] 5) Process and update the model in real time in the browser, ensuring that users can interact and observe the fitted structure;

[0065] 6) The above steps can use the least squares method to fit a plane to the point cloud data, so as to obtain the optimal plane parameters and minimize the error from the point to the plane;

[0066] 7) After the model is fitted in the browser, the fitting results can be fed back and updated in real time. Users can change the point cloud data through interaction to observe the changes in the fitting effect.

[0067] As a preferred technical solution of the present invention, the cloud computing support specific manner is as follows:

[0068] 1) Design a distributed architecture that combines cloud computing resources and browser user structure. The architecture includes the following components:

[0069] Cloud computing resources: Use computing instances from cloud service providers (such as AWS, Azure, and Google Cloud) to ensure sufficient computing power to process large-scale point cloud data;

[0070] Backend service: provides API interfaces for data processing and analysis, accepts requests from browsers, and distributes tasks to cloud computing resources for execution;

[0071] Front-end interface: Provides a user-friendly browser interface for users to upload, manage and monitor point cloud data processing tasks;

[0072] 2) Complete data uploading and preprocessing. The specific sub-steps are as follows:

[0073] Sub-step 1: Users upload large-scale point cloud data files through the browser interface, using modern front-end technologies (such as Dropzone.js) or directly uploading to the back-end service through the Web API;

[0074] Sub-step 2: The backend service receives the uploaded data and performs necessary preprocessing operations, such as common data format conversion, denoising filtering, and downsampling. The above operations are aimed at preparing the data for subsequent computing tasks;

[0075] 3) Computing task management and scheduling, the specific sub-steps are as follows:

[0076] Sub-step 1: The user submits a computing task through the browser interface, selects the analysis method and parameters, and the backend service receives the task request and manages the task queue;

[0077] Sub-step 2: The backend service distributes the task to cloud computing resources for execution based on the task type and computing requirements. These resources may include virtual machine instances, containers, or specialized computing services;

[0078] Sub-step 3: The browser interface can display the task progress and calculation results in real time. Use WebSocket technology for real-time updates, so that users can instantly obtain task status and results;

[0079] 4) Data analysis and visualization, the specific sub-steps are as follows:

[0080] Sub-step 1: Execute the point cloud data analysis task on cloud computing resources;

[0081] Sub-step 2: Return the analysis results to the browser and visualize the processed point cloud data through WebGL or other front-end 3D libraries.

[0082] As a preferred technical solution of the present invention, the specific process of the interactive analysis and visualization processing is as follows:

[0083] 1) Users can upload point cloud data files through the interface, which can support drag-and-drop upload or file selection;

[0084] 2) The backend receives the uploaded data and performs data preprocessing in various pre-processing steps;

[0085] 3) Users can select the analysis tasks to be performed, such as feature extraction, geometric analysis, and point cloud registration, and the interface needs to provide clear options and parameter settings;

[0086] 4) The backend distributes tasks to appropriate computing resources for centralized processing based on the task type selected by the user;

[0087] 5) Use WebSocket or polling technology to display the analysis process and progress in real time on the interface. This can be a progress bar, log update or dynamic chart;

[0088] 6) Present the processed point cloud data to the user in the form of 3D scenes or charts; use WebGL or similar technologies to achieve high-performance 3D visualization effects;

[0089] 7) Users can choose to export processed data in multiple formats such as PLY, OBJ, etc. The interface should provide options for selecting export areas or specific analysis results;

[0090] 8) Provide users with the ability to generate sharing links or invite other users to view specific data or analysis results.

[0091] Compared with the prior art, the present invention provides a browser-based online solution and processing method for laser point cloud data, which has the following beneficial effects:

[0092] This browser-based online solution and processing method for laser point cloud data avoids the cumbersome installation and upgrade procedures of traditional clients and the trouble of replacing equipment.

[0093] At the same time, the processing method realizes real-time display and interaction. It can realize the real-time user's rotation and zooming of the scene through view transformation and camera control, so that the operation can change the viewing angle. At the same time, it realizes the point cloud data selection and selection function, and can detect the point cloud information clicked by the user through ray picking.

[0094] This method can effectively record a large amount of laser point cloud data into a modern Web browser, and complete the display and interaction in real time, providing users with effective point cloud data analysis and visualization experience. At the same time, the architecture and process design involved in this method can realize the efficient processing and analysis of large-scale point cloud data, and realize the management and scheduling of tasks on the browser interface, meeting the user's real-time interactive needs for multi-dimensional data processing, and supporting data export and sharing functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 The present invention is a schematic diagram of the implementation process of the browser-based online solution and processing method of laser point cloud data. DETAILED DESCRIPTION

[0096] The following will be combined with the embodiments of the present invention and the accompanying drawings to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0097] See also Figure 1 In this embodiment: a browser-based online solution and processing method for laser point cloud data includes the following steps:

[0098] Step 1: Data loading and display:

[0099] Use WebGL technology to load a large amount of laser point cloud data into the browser, and display and interact in real time through 3D scene rendering technology;

[0100] The specific steps for data loading and display are as follows:

[0101] 1) Laser point cloud data is usually stored in LAS (Lidar Data Exchange Format) or PTS (Point Cloud Data) format. Before loading into WebGL, the following preprocessing may be required:

[0102] Data format conversion: Convert the original point cloud data format into a format that can be processed by WebGL, usually using JSON and binary data;

[0103] Data compression: compress data to reduce transmission and preloading time;

[0104] Coordinate system conversion: The coordinate system of the point cloud data can be converted into a coordinate system more suitable for WebGL rendering as needed;

[0105] 2) Efficiency in recording large amounts of point cloud data in WebGL is key, and there are usually several approaches:

[0106] Load directly into memory: For small-scale point cloud data, the data can be directly stored in JavaScript variables;

[0107] Block loading: For large-scale data, the block loading method is adopted to dynamically load and unload data blocks according to the field of view;

[0108] Using buffer objects: Buffer objects in WebGL manage and process large amounts of data at a constant and efficient speed. Through buffer mapping, point cloud data can be sent to the GPU for parallel processing and rendering.

[0109] 3) Since WebGL is a web graphics library based on OpenGL ES, it allows the use of JavaScript in web browsers for hardware-accelerated 3D graphics rendering. In point cloud data rendering, the following steps can be used to achieve it:

[0110] Vertex shader and fragment shader: The vertex shader is used to process the position and color information of the point, and the fragment shader is used to render the final color of each point;

[0111] Buffer Objects: Use buffers to manage vertex data and attribute data for efficient processing and rendering on the GPU;

[0112] Shader Programs: Write and optimize vertex shader and fragment shader programs to maximize rendering performance and quality;

[0113] 4) To complete real-time display and interaction, you can use the following methods:

[0114] View transformation and camera control: Enable users to rotate, zoom, and translate the scene, so that the operation can change the viewing angle;

[0115] Point selection and performance: realize the point selection and selection function of point cloud data, and detect the point cloud clicked by the user through ray picking;

[0116] Dynamic update: dynamically update the rendered point cloud data based on user interaction and data updates;

[0117] 5) When loading and rendering point cloud data, performance and security issues also need to be considered:

[0118] Data flow and segmented loading: Avoid recording too much data at one time. Use appropriate data flow and segmented loading strategies to improve rendering efficiency and user experience.

[0119] Ensure data is transmitted via HTTPS to prevent sensitive data from being leaked during transmission;

[0120] Step 2: Data preprocessing and filtering:

[0121] Implement preprocessing of laser point cloud data in the browser, such as denoising, filtering and registration operations;

[0122] The denoising and filtering processes are as follows:

[0123] 1) First, ensure that the laser point cloud data can be effectively loaded and parsed in WebGL;

[0124] 2) Select an effective denoising and filtering algorithm, which can be completed by using a statistical filter, a voxel grid filter, or a Gaussian filter;

[0125] 3) To implement the above algorithm in a browser environment, an efficient JavaScript code is usually required, which is completed by using a statistical filter. The specific code is as follows:

[0126]

[0127]

[0128] 4) During the implementation process, debugging and optimization are required to ensure that the algorithm can be processed in the browser within a reasonable time;

[0129] The registration process is as follows:

[0130] 1) It is necessary to ensure that two sets of point cloud data can be loaded and parsed, used as source point cloud and target point cloud respectively, and the data can be obtained through file loading or network request;

[0131] 2) Use the iterative closest point algorithm to find the optimal rotation and translation transformation so that the two sets of point clouds overlap as much as possible. The ICP algorithm JavaScript code is as follows:

[0132]

[0133]

[0134] 3) Then use the appropriate WebGL library to render the processed point cloud data to the browser so that users can view the registration results. During the above process, debugging and performance optimization are required;

[0135] Step 3: Feature data extraction and effective analysis:

[0136] Browser-based algorithms enable feature data extraction from point clouds, such as ground extraction, object recognition and classification;

[0137] The ground extraction of point cloud data is as follows:

[0138] 1) Complete the loading of point cloud data by loading files;

[0139] 2) Random Sampling Consensus (RANSAC) is used to complete ground extraction. The JavaScript code is as follows:

[0140]

[0141]

[0142] 3) After extracting ground and non-ground points, further processing of this type of data is completed, and the processed point cloud data is rendered to the browser using WebGL;

[0143] The ground extraction of point cloud data is as follows:

[0144] 1) Complete the loading of point cloud data by loading files;

[0145] 2) Select the object segmentation method based on region generation to implement object segmentation in the browser. The JavaScript code is as follows:

[0146]

[0147]

[0148]

[0149] 3) After object recognition is completed, the identified object point cloud is processed and visualized;

[0150] The classification and processing of point cloud data are as follows:

[0151] 1) Complete the loading of point cloud data by loading files;

[0152] 2) Select the point cloud classification method based on support vector machine (SVM) to classify the point cloud data in the browser. The JavaScript code is as follows:

[0153]

[0154]

[0155]

[0156] 3) After completing the point cloud classification, the classification results can be processed and visualized;

[0157] Step 4: Real-time solution and model fitting:

[0158] Use the computing power in the browser to perform real-time settlement and model fitting of point cloud data;

[0159] The specific steps of real-time settlement and model fitting are as follows:

[0160] 1) Complete the loading of the pre-processed point cloud data in the above steps;

[0161] 2) For point cloud data, it is necessary to extract features and calculate descriptors. Feature extraction can be based on the geometric properties, normal vectors, and curvature of the point cloud, while the descriptor is a mathematical description of the local or global structure;

[0162] 3) Select the appropriate model according to the application scenario;

[0163] 4) Use a numerical optimization algorithm to fit the model to the point cloud data. The numerical optimization algorithm can be selected from the least squares method and the gradient descent method;

[0164] 5) Process and update the model in real time in the browser, ensuring that users can interact and observe the fitted structure;

[0165] 6) The above steps can use the least squares method to fit a plane to the point cloud data, so as to obtain the optimal plane parameters and minimize the error from the point to the plane. The JavaScript formula code is as follows:

[0166]

[0167]

[0168] 7) After the model is fitted in the browser, the fitting results can be fed back and updated in real time. Users can change the point cloud data interactively to observe the changes in the fitting effect;

[0169] Step 5: Complete cloud computing support:

[0170] Combined with cloud computing resources, it can realize efficient processing and analysis of large-scale point cloud data, and manage and schedule computing tasks through the browser interface;

[0171] Cloud computing supports the following specific methods:

[0172] 1) Design a distributed architecture that combines cloud computing resources and browser user structure. The architecture includes the following components:

[0173] Cloud computing resources: Use computing instances from cloud service providers (such as AWS, Azure, and Google Cloud) to ensure sufficient computing power to process large-scale point cloud data;

[0174] Backend service: provides API interfaces for data processing and analysis, accepts requests from browsers, and distributes tasks to cloud computing resources for execution;

[0175] Front-end interface: Provides a user-friendly browser interface for users to upload, manage and monitor point cloud data processing tasks;

[0176] 2) Complete data uploading and preprocessing. The specific sub-steps are as follows:

[0177] Sub-step 1: Users upload large-scale point cloud data files through the browser interface, using modern front-end technologies (such as Dropzone.js) or directly uploading to the back-end service through the Web API;

[0178] Sub-step 2: The backend service receives the uploaded data and performs necessary preprocessing operations, such as common data format conversion, denoising filtering, and downsampling. The above operations are aimed at preparing the data for subsequent computing tasks;

[0179] 3) Computing task management and scheduling, the specific sub-steps are as follows:

[0180] Sub-step 1: The user submits a computing task through the browser interface, selects the analysis method and parameters, and the backend service receives the task request and manages the task queue;

[0181] Sub-step 2: The backend service distributes the task to cloud computing resources for execution based on the task type and computing requirements. These resources may include virtual machine instances, containers, or specialized computing services;

[0182] Sub-step 3: The browser interface can display the task progress and calculation results in real time. Use WebSocket technology for real-time updates, so that users can instantly obtain task status and results;

[0183] 4) Data analysis and visualization, the specific sub-steps are as follows:

[0184] Sub-step 1: Execute the point cloud data analysis task on cloud computing resources;

[0185] Sub-step 2: Return the analysis results to the browser and visualize the processed point cloud data through WebGL or other front-end 3D libraries;

[0186] Step 6: Complete interactive analysis and visualization:

[0187] Provides a user-friendly interactive interface that allows users to perform multi-dimensional analysis and real-time visualization of point cloud data in the browser, and supports data export and sharing;

[0188] The specific process of interactive analysis and visualization is as follows:

[0189] 1) Users can upload point cloud data files through the interface, which can support drag-and-drop upload or file selection;

[0190] 2) The backend receives the uploaded data and performs data preprocessing in various pre-processing steps;

[0191] 3) Users can select the analysis tasks to be performed, such as feature extraction, geometric analysis, and point cloud registration, and the interface needs to provide clear options and parameter settings;

[0192] 4) The backend distributes tasks to appropriate computing resources for centralized processing based on the task type selected by the user;

[0193] 5) Use WebSocket or polling technology to display the analysis process and progress in real time on the interface. This can be a progress bar, log update or dynamic chart;

[0194] 6) Present the processed point cloud data to the user in the form of 3D scenes or charts; use WebGL or similar technologies to achieve high-performance 3D visualization effects;

[0195] 7) Users can choose to export processed data in multiple formats such as PLY, OBJ, etc. The interface should provide options for selecting export areas or specific analysis results;

[0196] 8) Provide users with the ability to generate sharing links or invite other users to view specific data or analysis results.

[0197] In the above embodiments, it should be understood that the disclosed algorithms and systems can be implemented in other ways; for example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed; another point, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms;

[0198] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0199] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A browser-based online solution and processing method for laser point cloud data, characterized in that: The steps include: Step 1: Data loading and display: Use WebGL technology to load a large amount of laser point cloud data into the browser, and display and interact in real time through 3D scene rendering technology; Step 2: Data preprocessing and filtering: Implement preprocessing of laser point cloud data in the browser, such as denoising, filtering and registration operations; Step 3: Feature data extraction and effective analysis: Browser-based algorithms enable feature data extraction from point clouds, such as ground extraction, object recognition and classification; Step 4: Real-time solution and model fitting: Use the computing power in the browser to perform real-time settlement and model fitting of point cloud data; Step 5: Complete cloud computing support: Combined with cloud computing resources, it can realize efficient processing and analysis of large-scale point cloud data, and manage and schedule computing tasks through the browser interface; Step 6: Complete interactive analysis and visualization: It provides a user-friendly interactive interface, allowing users to perform multi-dimensional analysis and real-time visualization of point cloud data in the browser, and supports data export and sharing.

2. The browser-based online laser point cloud data calculation and processing method according to claim 1, characterized in that: The specific steps for loading and displaying data in step 1 are as follows: 1) Laser point cloud data is usually stored in LAS (Lidar Data Exchange Format) or PTS (Point Cloud Data) format. Before loading into WebGL, the following preprocessing may be required: Data format conversion: Convert the original point cloud data format into a format that can be processed by WebGL, usually using JSON and binary data; Data compression: compress data to reduce transmission and preloading time; Coordinate system conversion: The coordinate system of the point cloud data can be converted into a coordinate system more suitable for WebGL rendering as needed; 2) Efficiency in recording large amounts of point cloud data in WebGL is key, and there are usually several approaches: Load directly into memory: For small-scale point cloud data, the data can be directly stored in JavaScript variables; Block loading: For large-scale data, the block loading method is adopted to dynamically load and unload data blocks according to the field of view; Using buffer objects: Buffer objects in WebGL manage and process large amounts of data at a constant and efficient speed. Through buffer mapping, point cloud data can be sent to the GPU for parallel processing and rendering. 3) Since WebGL is a web graphics library based on OpenGL ES, it allows the use of JavaScript in web browsers for hardware-accelerated 3D graphics rendering. In point cloud data rendering, the following steps can be used to achieve it: Vertex shader and fragment shader: The vertex shader is used to process the position and color information of the point, and the fragment shader is used to render the final color of each point; Buffer Objects: Use buffers to manage vertex data and attribute data for efficient processing and rendering on the GPU; Shader Programs: Write and optimize vertex shader and fragment shader programs to maximize rendering performance and quality; 4) To complete real-time display and interaction, you can use the following methods: View transformation and camera control: Enable users to rotate, zoom, and translate the scene, so that the operation can change the viewing angle; Point selection and performance: realize the point selection and selection function of point cloud data, and detect the point cloud clicked by the user through ray picking; Dynamic update: dynamically update the rendered point cloud data based on user interaction and data updates; 5) When loading and rendering point cloud data, performance and security issues also need to be considered: Data flow and segmented loading: Avoid recording too much data at one time. Use appropriate data flow and segmented loading strategies to improve rendering efficiency and user experience. Ensure data is transmitted via HTTPS to prevent sensitive data from being leaked during transmission.

3. The browser-based online laser point cloud data solution and processing method according to claim 1, characterized in that: The denoising and filtering process described in step 2 is as follows: 1) First, ensure that the laser point cloud data can be effectively loaded and parsed in WebGL; 2) Select an effective denoising and filtering algorithm, which can be completed by using a statistical filter, a voxel grid filter, or a Gaussian filter; 3) To implement the above algorithm in a browser environment, an efficient JavaScript code is usually required, which is completed by using a statistical filter; 4) During the implementation process, debugging and optimization are required to ensure that the algorithm can be processed in the browser within a reasonable time; The registration process is specifically as follows: 1) It is necessary to ensure that two sets of point cloud data can be loaded and parsed, used as source point cloud and target point cloud respectively, and the data can be obtained through file loading or network request; 2) Use the iterative closest point algorithm to find the optimal rotation and translation transformation so that the two sets of point clouds overlap as much as possible; 3) Then use the appropriate WebGL library to render the processed point cloud data into the browser so that users can view the alignment results. In the above process, debugging and performance optimization are required.

4. The browser-based online laser point cloud data calculation and processing method according to claim 1, characterized in that: The ground extraction of point cloud data in step 3 is as follows: 1) Complete the loading of point cloud data by loading files; 2) Random Sampling Consensus (RANSAC) is used to complete ground extraction; The ground extraction of point cloud data is specifically as follows: 1) Complete the loading of point cloud data by loading files; 2) Select the object segmentation method based on region production and implement object segmentation in the browser; 3) After object recognition is completed, the identified object point cloud is processed and visualized; The point cloud data classification process is specifically as follows: 1) Complete the loading of point cloud data by loading files; 2) Select the point cloud classification method based on support vector machine (SVM) to classify the point cloud data in the browser; 3) After completing the point cloud classification, the classification results can be processed and visualized.

5. The browser-based online laser point cloud data solution and processing method according to claim 1, characterized in that: The real-time settlement and model fitting steps described in step 4 are as follows: 1) Complete the loading of the pre-processed point cloud data in the above steps; 2) For point cloud data, it is necessary to extract features and calculate descriptors. Feature extraction can be based on the geometric properties, normal vectors, and curvature of the point cloud, while the descriptor is a mathematical description of the local or global structure; 3) Select the appropriate model according to the application scenario; 4) Using a numerical optimization algorithm to fit the model to the point cloud data, the numerical optimization algorithm may be one of the least squares method and the gradient descent method; 5) Process and update the model in real time in the browser, ensuring that users can interact and observe the fitted structure; 6) The above steps can use the least squares method to fit a plane to the point cloud data, so as to obtain the optimal plane parameters and minimize the error from the point to the plane; 7) After the model is fitted in the browser, the fitting results can be fed back and updated in real time. Users can change the point cloud data through interaction to observe the changes in the fitting effect.

6. The browser-based online laser point cloud data solution and processing method according to claim 1, characterized in that: The specific methods of cloud computing support described in step 5 are as follows: 1) Design a distributed architecture that combines cloud computing resources and browser user structure. The architecture includes the following components: Cloud computing resources: Use computing instances from cloud service providers (e.g. AWS, Azure, Google CLOUD) to ensure sufficient computing power to process large-scale point cloud data; Backend service: provides API interfaces for data processing and analysis, accepts requests from browsers, and distributes tasks to cloud computing resources for execution; Front-end interface: Provides a user-friendly browser interface for users to upload, manage and monitor point cloud data processing tasks; 2) Complete data uploading and preprocessing. The specific sub-steps are as follows: Sub-step 1: Users upload large-scale point cloud data files through the browser interface, using modern front-end technologies (such as Dropzone.js) or directly uploading to the back-end service through the Web API; Sub-step 2: The backend service receives the uploaded data and performs necessary preprocessing operations, such as common data format conversion, denoising filtering, and downsampling. The above operations are aimed at preparing the data for subsequent computing tasks; 3) Computing task management and scheduling, the specific sub-steps are as follows: Sub-step 1: The user submits a computing task through the browser interface, selects the analysis method and parameters, and the backend service receives the task request and manages the task queue; Sub-step 2: The backend service distributes the task to cloud computing resources for execution based on the task type and computing requirements. These resources may include virtual machine instances, containers, or specialized computing services; Sub-step 3: The browser interface can display the task progress and calculation results in real time. Use WebSocket technology for real-time updates, so that users can instantly obtain task status and results; 4) Data analysis and visualization, the specific sub-steps are as follows: Sub-step 1: Execute the point cloud data analysis task on cloud computing resources; Sub-step 2: Return the analysis results to the browser and visualize the processed point cloud data through WebGL or other front-end 3D libraries.

7. The browser-based online laser point cloud data solution and processing method according to claim 1, characterized in that: The specific process of interactive analysis and visualization processing described in step 6 is as follows: 1) Users can upload point cloud data files through the interface, which can support drag-and-drop upload or file selection; 2) The backend receives the uploaded data and performs data preprocessing in various pre-processing steps; 3) Users can select the analysis tasks to be performed, such as feature extraction, geometric analysis, and point cloud registration, and the interface needs to provide clear options and parameter settings; 4) The backend distributes tasks to appropriate computing resources for centralized processing based on the task type selected by the user; 5) Use WebSocket or polling technology to display the analysis process and progress in real time on the interface. This can be a progress bar, log update or dynamic chart; 6) Present the processed point cloud data to the user in the form of 3D scenes or charts; use WebGL or similar technologies to achieve high-performance 3D visualization effects; 7) Users can choose to export processed data in multiple formats such as PLY, OBJ, etc. The interface needs to provide options for selecting export areas or specific analysis results; 8) Provide users with the ability to generate sharing links or invite other users to view specific data or analysis results.

Images