A multi-terminal-oriented point cloud data format standardization conversion method

Abstract

The application discloses a multi-terminal-oriented point cloud data format standardization conversion method and relates to the technical field of three-dimensional laser scanning, which comprises the following steps: performing downsampling processing on block point cloud data to generate multi-level resolution point cloud representation; performing a dynamic dilution operation on multi-resolution point cloud data to obtain dilution point cloud data files; performing format conversion on the dilution point cloud data by using a plug-in modular method to export point cloud data format files suitable for multiple platforms; dynamically adjusting the resolution and block range of the point cloud data according to the real-time loading block point cloud data, and outputting optimized point cloud data by loading new plug-ins to adapt to new point cloud formats; and combining the characteristics of a substation scene and point cloud geometry, efficiently segmenting and diluting the point cloud data according to the importance score of points, so that the data volume is greatly reduced on the basis of ensuring the integrity of the information of key areas.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of three-dimensional laser scanning, in particular to a multi-terminal-oriented point cloud data format standardization conversion method. BACKGROUND

[0002] In recent years, with the rapid development of three-dimensional laser scanning technology, point cloud data has been widely used in digital modeling of substations, equipment state monitoring and intelligent operation and maintenance. The current point cloud acquisition technology usually relies on fixed scanning equipment to generate point cloud data through multi-angle scanning and post-processing. However, due to the complex structure of the substation scene and the irregular distribution of equipment, the traditional method has certain limitations in acquisition efficiency, data integrity and subsequent processing convenience. In addition, the demand for multi-terminal adaptation of point cloud data is increasingly prominent, especially in real-time rendering and intelligent decision-making scenarios, different terminals have different requirements for the resolution, file size and loading speed of point cloud data.

[0003] At present, point cloud data processing technology has made certain progress in denoising, blocking and compression, but still has the following shortcomings: first, the existing denoising algorithm lacks pertinence when removing invalid points, and effective information in key areas is easily lost; second, the existing blocking method relies on fixed parameter settings and fails to fully combine the spatial characteristics of point cloud data, resulting in inefficient blocking results; third, the multi-terminal adaptation scheme lacks dynamic adjustment capability and cannot flexibly optimize point cloud resolution and loading speed according to terminal performance and user demand. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a multi-terminal-oriented point cloud data format standardization conversion method to solve the problems of low efficiency and insufficient adaptability in the process of point cloud data acquisition, processing and multi-terminal adaptation in complex substation scenes.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a multi-terminal-oriented point cloud data format standardization conversion method, which includes collecting three-dimensional space point cloud data and preprocessing to generate a standardized point cloud data file;

[0008] The standardized point cloud data is processed by a spatial segmentation algorithm to output structured blocked point cloud data;

[0009] The blocked point cloud data is down-sampled to generate a multi-level resolution point cloud representation;

[0010] Perform a dynamic dilution operation on the multi-resolution point cloud data to obtain a dilution point cloud data file;

[0011] Convert the dilution point cloud data into a format using a plug-in modular method, and export a point cloud data format file adapted to multiple platforms;

[0012] Deploy the point cloud data format file to a multi-terminal device, dynamically load the block data according to the terminal performance and user demand, and output the real-time loaded block point cloud data;

[0013] According to the real-time loaded block point cloud data, dynamically adjust the resolution and block range of the point cloud data, and load new plug-ins to adapt to new point cloud formats, and output optimized point cloud data.

[0014] As a preferred scheme of the multi-terminal point cloud data format standardization conversion method, wherein: collecting three-dimensional space point cloud data and preprocessing to generate a standardized point cloud data file includes the following steps,

[0015] According to the complexity of the structure of the substation, a scanning path is planned;

[0016] Use a handheld laser radar scanning device to scan along the planned path and collect spatial point cloud data in real time;

[0017] Real-time positioning of the position and preliminary splicing of the point cloud data are performed through a SLAM algorithm, and an original point cloud data file in a local coordinate system is output;

[0018] The point cloud data in the original point cloud data file is denoised;

[0019] In combination with the scene characteristics of the substation, the denoised point cloud data is removed from the invalid points;

[0020] The point cloud data after removing the invalid points is converted from the local coordinate system to the global coordinate system;

[0021] For the point cloud data in the global coordinate system, the point cloud density is uniformized using a voxel grid method;

[0022] The uniformized point cloud data is supplemented and corrected in color to generate a standardized point cloud data file.

[0023] As a preferred scheme of the multi-terminal point cloud data format standardization conversion method, wherein: the standardized point cloud data is processed by a spatial segmentation algorithm, and structured block point cloud data is output, including the following steps,

[0024] The standardized point cloud data file is analyzed in terms of spatial range, the overall bounding box of the point cloud data is determined, and the segmentation parameters are initialized;

[0025] determine the maximum partition depth of the octree according to the density of the point cloud data;

[0026] Based on the overall bounding box of the point cloud data and the maximum partition depth, the point cloud data is recursively partitioned using the octree algorithm, and a unique spatial index number is generated for each partition;

[0027] An index table is established for the generated partitioned point cloud data, and the point cloud data with a unique spatial index number is stored as a structured partitioned point cloud data file according to the storage path recorded in the index table.

[0028] As a preferred scheme of the multi-terminal oriented point cloud data format standardization conversion method, the multi-resolution point cloud representation generated by downsampling the partitioned point cloud data includes the following steps,

[0029] According to the target application scenario and terminal performance requirements, the number of layers of multi-resolution and the downsampling ratio of each layer resolution are determined;

[0030] The original data of the partitioned point cloud is saved as a high-resolution hierarchical file in a standard format;

[0031] According to the voxel size of the medium resolution layer, the partitioned point cloud data in the high-resolution hierarchical file is divided into a three-dimensional voxel grid;

[0032] Aggregate each voxel in the three-dimensional voxel grid to generate medium-resolution hierarchical partitioned point cloud data;

[0033] Downsample the medium-resolution point cloud data of each partition again to generate a low-resolution hierarchical point cloud representation.

[0034] As a preferred scheme of the multi-terminal oriented point cloud data format standardization conversion method, the multi-resolution point cloud data is executed to obtain the diluted point cloud data file, including the following steps,

[0035] Load the multi-level resolution point cloud data of each partition;

[0036] Extract the geometric features of each level of resolution point cloud data through neighborhood analysis of the point cloud;

[0037] According to the semantic rules of the substation, the spatial position and attribute information of the point cloud are combined to score the importance of the point, and the expression is:

[0038]

[0039] Where S(P i ) represents the importance score of point P i , κ(P i) represents point P i The curvature, C(P) i ) represents point P i The normal vector, D(P) i ) represents point P i The local density, G(P) i ) represents point P i The attribute value, H(P) represents the average value of the attribute. i ) represents point P i The neighborhood entropy, σ(P) i ) represents point P i The standard deviation of the neighborhood attribute, P i This represents a point in the point cloud, where i represents the index of the point.

[0040] Based on the scoring results, distinguish between key areas and low-importance areas;

[0041] Based on geometric features and discrimination results, a dynamic dilution rule is designed to dynamically adjust the dilution ratio. The expression is:

[0042]

[0043] Among them, R(P) i ) represents point P i The dilution retention probability is denoted by α, which controls the dilution intensity in the transition region, β, which controls the decay rate of the dilution probability in the transition region, γ, which controls the dilution probability in the low importance region, T1, which represents the high threshold of the importance score, and T2, which represents the low threshold of the importance score.

[0044] Dynamic dilution operation is performed on the point cloud data according to the dynamic dilution rule. Point sets in important regions are directly retained, while point sets in secondary regions are randomly sampled according to the target dilution ratio.

[0045] The important region point sets and the diluted minor region point sets are merged to obtain the diluted point cloud data file.

[0046] As a preferred embodiment of the point cloud data format standardization conversion method for multiple terminals described in this invention, the method of converting diluted point cloud data using a plug-in modular approach to export point cloud data format files adapted to multiple platforms includes the following steps:

[0047] Read the local coordinates of each point from the diluted point cloud data file;

[0048] According to the requirements of the target 3D design platform, obtain the definition of the target's global coordinate system;

[0049] Use a linear transformation formula to convert local coordinates to global coordinates;

[0050] replacing the converted global coordinates with local coordinates in the point cloud file;

[0051] compressing the point cloud data file according to the corresponding compression standard using a mainstream point cloud compression library, quantizing the floating-point coordinates into integers, and converting the file format;

[0052] Based on the converted file, check whether it meets the compatibility requirements of the target platform, and verify whether the compressed point cloud can be correctly decoded and loaded to obtain the compressed point cloud file;

[0053] According to the point cloud format specification of the target platform, the attribute field of the compressed point cloud file is mapped to the compatible field of the target platform, and the point cloud data format file suitable for multiple platforms is exported.

[0054] As a preferred scheme of the multi-terminal oriented point cloud data format standardization conversion method, the point cloud data format file is deployed to a multi-terminal device, and the block data is dynamically loaded according to the terminal performance and user demand, and the real-time loaded block point cloud data is output, including the following steps,

[0055] The point cloud data format file is classified and deployed to the corresponding terminal device according to the data resolution level and terminal demand;

[0056] According to the user view range, the point cloud block to be loaded is dynamically determined;

[0057] According to the type of terminal device, the corresponding resolution level is selected;

[0058] The required point cloud data file is loaded block by block and stored in the memory, and the blocks that have been loaded but are beyond the user view range are dynamically unloaded to release the memory resources, and the real-time loaded block point cloud data is output.

[0059] As a preferred scheme of the multi-terminal oriented point cloud data format standardization conversion method, according to the real-time loaded block point cloud data, the resolution and block range of the point cloud data are dynamically adjusted, and the newly added point cloud format is adapted by loading a new plug-in, and the optimized point cloud data is output, including the following steps,

[0060] Capture the real-time operation of the user on the multi-terminal device and update the interaction state;

[0061] According to the user operation and the viewpoint distance, the point cloud resolution level is dynamically adjusted;

[0062] According to the resolution level and the view range, the newly added block data is retrieved from the index table, and the plug-in of the newly added point cloud format is dynamically loaded through the modular plug-in mechanism, and the optimized point cloud data is output.

[0063] In a second aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program, when executed by the processor, implements any step of the multi-terminal-oriented point cloud data format standardization conversion method according to the first aspect of the present application.

[0064] In a third aspect, the present application provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any step of the multi-terminal-oriented point cloud data format standardization conversion method according to the first aspect of the present application.

[0065] The present application has the following beneficial effects: by planning a scanning path and using a SLAM algorithm for real-time positioning and preliminary splicing, the scanning problem caused by the complexity of the substation structure is effectively solved; by point cloud denoising, invalid point elimination and global coordinate conversion, the quality of data processing is further improved, and the generated point cloud data is more standardized and structured; by using voxel grid uniformization and multi-level resolution downsampling technology, the point cloud data is more efficient in storage and transmission; octree blocking and dynamic dilution algorithm, combined with the characteristics of the substation scene and the geometric features of the point cloud, can efficiently segment and dilute the point cloud data according to the importance score of the points, thereby greatly reducing the data volume while ensuring the integrity of the information in the key areas; by using the modular plug-in mechanism, the standardized conversion of the point cloud format and the multi-terminal compatible deployment are realized, so that the point cloud data can adapt to the needs of different platforms and support real-time loading and dynamic adjustment. BRIEF DESCRIPTION OF DRAWINGS

[0066] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0067] Figure 1 The flowchart of the multi-terminal-oriented point cloud data format standardization conversion method in embodiment 1.

[0068] Figure 2 The schematic diagram of the multi-level resolution point cloud representation in embodiment 1. DETAILED DESCRIPTION

[0069] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification.

[0070] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0071] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0072] Example 1, referring to Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides a point cloud data format standardization conversion method for multiple terminals, including the following steps:

[0073] S1. Collect 3D spatial point cloud data and preprocess it to generate standardized point cloud data files.

[0074] S1.1. Based on the structural complexity of the substation, plan the scanning path; use a handheld lidar scanning device to scan along the planned path and collect spatial point cloud data in real time; use the SLAM algorithm to complete the real-time positioning of its own position and the initial stitching of the point cloud data, and output the original point cloud data file in the local coordinate system; perform noise reduction processing on the point cloud data in the original point cloud data file.

[0075] Specifically, the denoising process employs a statistical analysis-based denoising algorithm, and the specific operations are as follows:

[0076] Define a neighborhood radius *r* for the point cloud, and select neighborhood points within a fixed radius range (recommended value: *r* = 0.05–0.1 meters, depending on the point cloud density of the acquisition environment); for any point in the point cloud, count the number of points in its neighborhood, and calculate the deviation of that point from the mean of its neighborhood points, expressed as:

[0077]

[0078] Where, d i N represents the deviation at the i-th point. i This represents the number of neighboring points of the i-th point, within a set neighborhood radius r, at a distance from point p. i The number of points less than r, where j represents the index p of a point in the neighborhood. j Let j represent the j-th neighboring point.

[0079] like Then point p i Removed, of which σd is the standard deviation of the global bias d, is the global mean.

[0080] S1.2. According to the scene characteristics of the substation, the invalid points in the point cloud data after denoising are removed; the point cloud data after removing the invalid points is converted from the local coordinate system to the global coordinate system; for the point cloud data in the global coordinate system, the voxel grid method is used to uniformize the point cloud density; the color of the uniformized point cloud data is supplemented and corrected to generate a standardized point cloud data file.

[0081] Specifically, the isolated point removal is based on a point cloud density isolated point detection algorithm. For a point in the point cloud whose neighborhood points are less than a set threshold T min ( e.g. T min = 5), it is marked as an isolated point and removed.

[0082] If N i < T min , the point p i is removed.

[0083] The conversion from the local coordinate system to the global coordinate system is realized by calculating the conversion matrix of the acquisition device coordinate system and the global coordinate system through the global positioning reference points (such as GPS reference points or fixed calibration points) arranged in the substation, and the expression is:

[0084]

[0085] wherein T represents the coordinate conversion matrix, R represents the rotation matrix, and t represents the translation vector.

[0086] The local coordinates of each point are converted, and the expression is:

[0087]

[0088] wherein p i represents the three-dimensional point cloud coordinates of the converted point p in the global coordinate system, represents the original coordinates of the point p i in the local coordinate system.

[0089] The specific operation of point cloud density uniformization is as follows:

[0090] Voxel division: define the resolution s of the voxel grid (recommended value: s = 0.01-0.1 meters, depending on the accuracy requirements of the application scenario), and divide the three-dimensional space into voxels with a size of s x s x s.

[0091] Point cloud dilution: for each voxel, only one representative point (such as the voxel center point or the average value of the points in the voxel) is retained, and other points are removed, thereby realizing the uniformization of the point cloud density.

[0092] Point cloud data supplement and correction color is if the laser radar does not collect color information, combined with high resolution panoramic image, add color attribute to each point by projection mapping method.

[0093] S2. The standardized point cloud data is processed by a spatial segmentation algorithm, and structured segmented point cloud data is output.

[0094] S2.1. Analyze the spatial range of the standardized point cloud data file, determine the overall bounding box of the point cloud data, and initialize the segmentation parameters.

[0095] Further, the analysis of the spatial range is to express the three-dimensional spatial range of the point cloud data by the maximum and minimum coordinate values.

[0096] S2.2. Determine the maximum segmentation depth of the octree according to the density of the point cloud data.

[0097] Wherein, the density of the point cloud data can be defined as the number of points in a unit volume, and the density is usually analyzed by counting the distribution of points in the point cloud.

[0098] The volume of the segmentation unit of the maximum segmentation depth should be matched with the local point density as much as possible. Too deep segmentation will result in too small segmentation, increasing the storage and indexing burden; too shallow segmentation may not capture local details.

[0099] S2.3. Based on the overall bounding box of the point cloud data and the maximum segmentation depth, the point cloud data is segmented recursively using the octree algorithm, and a unique spatial index number is generated for each segment; an index table is established for the generated segmented point cloud data, and the point cloud data with a unique spatial index number is stored as a structured segmented point cloud data file according to the storage path recorded in the index table.

[0100] Specifically, the octree is a recursive spatial segmentation algorithm that divides a three-dimensional space into 8 subspaces (i.e. 8 sub-blocks). Each sub-block will continue to be recursively divided after each division until the maximum segmentation depth is reached or other termination conditions are met.

[0101] Recursive operation:

[0102] Divide the spatial range of each node into 2 along the coordinate axis to generate 8 child nodes; traverse the point cloud data and assign each point to the corresponding child node; if the segmentation depth of the node has reached the termination condition, stop the segmentation, otherwise continue to recursively segment the child nodes.

[0103] S3. The segmented point cloud data is down-sampled to generate multi-level resolution point cloud representation.

[0104] S3.1. Determine the number of layers and the down-sampling ratio of each layer according to the target application scenario and terminal performance requirements.

[0105] Specifically, the common resolution layer number:

[0106] High-resolution layer: complete original point cloud data, retaining all details.

[0107] Medium-resolution layer: generated by down-sampling, retaining main geometric features.

[0108] Low-resolution layer: further down-sampling, retaining only rough global structure.

[0109] S3.2. Save the original data of the block point cloud in a standard format as a high-resolution hierarchical file; divide the block point cloud data in the high-resolution hierarchical file into a three-dimensional voxel grid according to the voxel size of the medium-resolution layer; aggregate each voxel in the three-dimensional voxel grid to generate medium-resolution hierarchical block point cloud data; and down-sample the medium-resolution point cloud data of each block again to generate low-resolution hierarchical point cloud representation.

[0110] Specifically, the definition of voxel grid: the voxel grid organizes point cloud data by dividing three-dimensional space into fixed-size cubic units (called voxels).

[0111] The size of the voxel is determined by the down-sampling ratio of the medium-resolution layer.

[0112] Voxel aggregation: for each voxel, aggregate the points within the voxel into a representative point to generate medium-resolution layer point cloud.

[0113] S4. Perform dynamic thinning operation on the multi-resolution point cloud data to obtain thinned point cloud data file.

[0114] S4.1. Load the multi-layer resolution point cloud data of each block; extract the geometric features of each layer of point cloud data through neighborhood analysis of the point cloud; and score the importance of the points according to the semantic rules of the substation, combined with the spatial position and attribute information of the point cloud, the expression is:

[0115]

[0116] where S(P i ) represents the importance score of point P i , κ(P i ) represents the curvature of point P i , C(P i ) represents the consistency of the normal vector of point P i , D(P i ) represents the local density of point P i , and G(Pi represents the attribute value of point P i represents the average value of the attribute, H(P i ) represents the neighborhood entropy of point P i i ) represents the neighborhood attribute standard deviation of point P i i represents the point in the point cloud, and i represents the index of the point.

[0117] The expression of curvature is:

[0118] where λ max represents the maximum eigenvalue of the neighborhood point covariance matrix, and λ j represents the eigenvalue of the jth neighborhood point.

[0119] The expression of the consistency of the normal vector is:

[0120] where k represents the number of neighborhood points, n i represents the normal vector of the ith point, and n j represents the normal vector of the jth neighborhood point.

[0121] The expression of local density is:

[0122] The expression of neighborhood entropy is:

[0123] The expression of attribute standard deviation is:

[0124] S4.2. Based on the scoring results, distinguish the key area and the low importance area.

[0125] The area division is divided into three categories:

[0126] Key area: direct reservation; transition area: dynamic dilution; low importance area: strong dilution.

[0127] S4.3. According to the geometric features and the division results, design the dynamic dilution rule, and dynamically adjust the dilution ratio, the expression is:

[0128]

[0129] where R(P i ) represents the point P i ​​​wherein, D represents the dilution probability of the key region, a represents the dilution intensity of the transition region, β represents the decay rate of the dilution probability in the transition region, γ represents the dilution probability of the low importance region, T1 represents a high threshold value of the importance score for distinguishing the key region and the transition region, and T2 represents a low threshold value of the importance score for distinguishing the low importance region and the transition region;

[0130] According to the dynamic dilution rule, a dynamic dilution operation is performed on the point cloud data, the important region point set is directly reserved, and the secondary region point set is randomly sampled according to a target dilution ratio; the reserved important region point set and the diluted secondary region point set are merged to obtain a diluted point cloud data file.

[0131] It should be noted that the multi-resolution point cloud data is diluted by using the dynamic dilution rule. The diluted point cloud data file retains the detailed information of the key region, while reducing the number of points in the low importance region, thereby optimizing the storage and computing efficiency and meeting the application requirements of substations and other specific fields.

[0132] S5. Utilize the plug-in modular method to perform format conversion on the diluted point cloud data, and export a point cloud data format file adapted to multiple platforms.

[0133] S5.1. Read the local coordinates of each point from the diluted point cloud file;

[0134] According to the requirements of the target three-dimensional design platform, the definition of the target global coordinate system is obtained; the local coordinates are converted into global coordinates using a linear transformation formula; the converted global coordinates are replaced with the local coordinates in the point cloud file; the mainstream point cloud compression library is used to compress the point cloud data file according to the corresponding compression standard, quantize the floating-point coordinates to integers, and convert the file format.

[0135] Quantization formula of floating-point coordinates:

[0136]

[0137] wherein, Q(x) represents the quantized integer coordinate value, x represents the original floating-point coordinate value of the point, round represents the rounding operation, x min represents the minimum value of the coordinate value in the point cloud data, and step represents the quantization step length, and the coordinate increment corresponding to each integer unit.

[0138] S5.2. Based on the converted file, check whether it meets the compatibility requirements of the target platform, and verify whether the compressed point cloud can be correctly decoded and loaded, to obtain a compressed point cloud file; according to the point cloud format specification of the target platform, map the attribute field of the compressed point cloud file to the compatible field of the target platform, and export a point cloud data format file adapted to multiple platforms.

[0139] It should be noted that after the point cloud data is converted and compressed, it is checked whether the file meets the technical specifications of the target platform. If the file does not meet the specifications of the target platform, it may not be loaded or used correctly. By checking, it is ensured that the point cloud data can be directly run on the target platform, reducing the workload of developers for repairing format problems in the later stage.

[0140] S6. Deploy the point cloud data format file to a plurality of terminal devices, dynamically load the chunked data according to terminal performance and user demand, and output the real-time loaded chunked point cloud data.

[0141] S6.1. According to the data resolution hierarchy and terminal demand, the point cloud data format file is classified and deployed to the corresponding terminal device; according to the user view range, the point cloud chunks that need to be loaded are dynamically determined.

[0142] Specifically, the resolution hierarchy refers to hierarchical processing of point cloud data, which is divided into different levels according to the density or accuracy of points:

[0143] High resolution level: complete point cloud data, highest point density, suitable for high-performance devices (such as desktop workstations).

[0144] Medium resolution level: sparse processing of point cloud data, reducing point density, suitable for medium-performance devices (such as laptops or tablets).

[0145] Low resolution level: further sparse, suitable for low-performance devices (such as mobile phones or embedded devices).

[0146] S6.2. According to the type of terminal device, select the corresponding resolution level; load the required point cloud data file block by block and store it in the memory, and dynamically unload the chunks that have been loaded but are outside the user view range, release the memory resources, and output the real-time loaded chunked point cloud data.

[0147] Specifically, the dynamic unloading mechanism refers to periodically checking the loaded chunked data in the memory, determining which chunks no longer overlap with the user view range, and preferentially unloading the chunks outside the view range. For larger chunk files, they can be unloaded in batches.

[0148] S7. According to the real-time loaded chunked point cloud data, dynamically adjust the resolution and chunk range of the point cloud data, and adapt to new point cloud formats by loading new plugins, and output the optimized point cloud data.

[0149] S7.1. Capture the real-time operation of the user on the multi-terminal device, update the interaction state; dynamically adjust the point cloud resolution level according to the user operation and the viewpoint distance; according to the resolution level and the view range, retrieve the newly added block data from the index table, and dynamically load the plug-in of the newly added point cloud format through the modular plug-in mechanism, and output the optimized point cloud data.

[0150] Specifically, the viewpoint distance refers to the distance between the camera position of the user and the point cloud data:

[0151] Close distance: the user needs to observe the details of the point cloud, and high-resolution data should be loaded.

[0152] Medium distance: medium-resolution data can be loaded to balance detail and performance.

[0153] Far distance: only low-resolution data needs to be loaded to reduce the number of points drawn.

[0154] The embodiment also provides a computer device suitable for the multi-terminal-oriented point cloud data format standardization conversion method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the multi-terminal-oriented point cloud data format standardization conversion method proposed in the above embodiment.

[0155] The computer device can be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved through WIFI, operator network, NFC (near field communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device. It can also be an external keyboard, touchpad or mouse, etc.

[0156] The embodiment also provides a storage medium, which stores a computer program, and the computer program is executed by a processor to implement the multi-terminal-oriented point cloud data format standardization conversion method proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk, or an optical disk.

[0157] To sum up, the application effectively solves the scanning problem caused by the complexity of the substation structure by planning a scanning path and using a SLAM algorithm for real-time positioning and preliminary splicing; the quality of data processing is further improved by point cloud denoising, invalid point elimination, and global coordinate conversion, ensuring that the generated point cloud data is more standardized and structured; the use of voxel grid uniformization and multi-level resolution downsampling technology makes the point cloud data more efficient in storage and transmission; the octree blocking and dynamic dilution algorithm, combined with the characteristics of the substation scene and the geometric features of the point cloud, can efficiently segment and dilute the point cloud data according to the importance score of the points, thereby greatly reducing the data volume while ensuring the integrity of the information in the key areas; the modular plug-in mechanism realizes the standardized conversion of the point cloud format and the multi-terminal compatible deployment, so that the point cloud data can adapt to the needs of different platforms and support real-time loading and dynamic adjustment.

[0158] It should be noted that the above embodiments are only used to illustrate the technical solutions of the application rather than limit the application. Although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the application, and all of them should be covered in the scope of the claims of the application.

Claims

1. A method for multi-terminal oriented point cloud data format standardization conversion, characterized in that: The application relates to a method for processing point cloud data of a substation. The method comprises the following steps: collecting three-dimensional space point cloud data and preprocessing to generate a standardized point cloud data file; using a space segmentation algorithm to block process the standardized point cloud data and output structured block point cloud data; performing down-sampling processing on the block point cloud data to generate multi-level resolution point cloud representation; performing a dynamic dilution operation on the multi-resolution point cloud data to obtain a dilution point cloud data file including the following steps: loading multi-level resolution point cloud data of each block; extracting geometric features of each level resolution point cloud data through neighborhood analysis of the point cloud; ； wherein, an importance score of the point a curvature of the point a normal vector of the point a local density of the point an attribute value of the point an average value of the attribute a neighborhood entropy of the point a neighborhood attribute standard deviation of the point representing a point in the point cloud, representing an index of the point;​​​​​​​ according to the semantic rules of the substation, combining the spatial position and attribute information of the point cloud to score the importance of the points, the expression being: based on the scoring results, distinguishing between key areas and low importance areas; ； wherein, a dilution retention probability, a dilution retention probability, a dilution strength controlling the transition region, a decay rate controlling the dilution probability in the transition region, a dilution probability of the low importance region, a high threshold value of the importance score, a low threshold value of the importance score; according to the geometric features and the distinguishing results, designing a dynamic dilution rule to dynamically adjust the dilution ratio, the expression being: performing a dynamic dilution operation on the point cloud data according to the dynamic dilution rule, directly retaining the important area point set, and randomly sampling the secondary area point set according to the target dilution ratio; merging the retained important area point set and the diluted secondary area point set to obtain a dilution point cloud data file; using a plug-in modular method to convert the dilution point cloud data into a format, and exporting a point cloud data format file that is adapted to multiple platforms; deploying the point cloud data format file to multiple terminal devices, dynamically loading block data according to terminal performance and user demand, and outputting real-time loaded block point cloud data; 2. The multi-terminal oriented point cloud data format standardization conversion method of claim 1, wherein: according to the real-time loaded block point cloud data, dynamically adjusting the resolution and block range of the point cloud data, and outputting optimized point cloud data by loading new plug-ins to adapt to new point cloud formats. The method comprises the following steps: collecting three-dimensional space point cloud data and preprocessing to generate a standardized point cloud data file including the following steps: planning a scanning path according to the structural complexity of the substation; scanning along the planned path using a handheld laser radar scanning device to collect space point cloud data in real time; performing real-time positioning of the self-position and preliminary splicing of the point cloud data through a SLAM algorithm, and outputting an original point cloud data file in a local coordinate system; performing denoising processing on the point cloud data in the original point cloud data file; combining the scene characteristics of the substation, removing invalid points from the denoised point cloud data; converting the point cloud data after removing invalid points from the local coordinate system to a global coordinate system; 3. The multi-terminal oriented point cloud data format standardization conversion method of claim 2, wherein: using a voxel grid method to uniformize the point cloud density for the point cloud data in the global coordinate system; supplementing and correcting the color of the uniformized point cloud data to generate a standardized point cloud data file. The method comprises the following steps: performing space range analysis on the standardized point cloud data file to determine the overall bounding box of the point cloud data and initialize segmentation parameters; determining the maximum segmentation depth of the octree according to the density of the point cloud data; based on the overall bounding box of the point cloud data and the maximum segmentation depth, using an octree algorithm to recursively block the point cloud data and generating a unique spatial index number for each block. An index table is established for the generated block point cloud data, and the point cloud data with a unique spatial index number is stored as a structured block point cloud data file according to the storage path recorded in the index table.

4. The multi-terminal oriented point cloud data format standardization conversion method of claim 3, wherein: The block point cloud data is down-sampled to generate a multi-level resolution point cloud representation, including the following steps, According to the target application scenario and terminal performance requirements, the number of layers of multi-level resolution and the down-sampling ratio of each layer resolution are determined; The original data of the block point cloud is saved in a standard format as a high-resolution hierarchical file; According to the voxel size of the medium resolution layer, the block point cloud data in the high-resolution hierarchical file is divided into a three-dimensional voxel grid; For each voxel in the three-dimensional voxel grid, aggregation is performed to generate medium-resolution hierarchical block point cloud data; The medium-resolution point cloud data of each block is down-sampled again to generate low-resolution hierarchical point cloud representation.

5. The multi-terminal oriented point cloud data format standardization conversion method of claim 4, wherein: The dilute point cloud data is converted into a format suitable for multiple platforms using a plug-in modular method, including the following steps, Read the local coordinates of each point from the dilute point cloud data file; According to the requirements of the target three-dimensional design platform, obtain the definition of the target global coordinate system; Convert the local coordinates to global coordinates using a linear transformation formula; Replace the local coordinates in the point cloud file with the converted global coordinates; Use the mainstream point cloud compression library to compress the point cloud data file according to the corresponding compression standard, quantize the floating-point coordinates to integers, and convert the file format; Based on the converted file, check whether it meets the compatibility requirements of the target platform, and verify whether the compressed point cloud can be correctly decoded and loaded, to obtain the compressed point cloud file; According to the point cloud format specification of the target platform, map the attribute fields of the compressed point cloud file to the compatible fields of the target platform, and export the point cloud data format file adapted to multiple platforms.

6. The multi-terminal oriented point cloud data format standardization conversion method of claim 5, wherein: Deploy the point cloud data format file to multiple terminal devices, dynamically load the block data according to the terminal performance and user demand, and output the real-time loaded block point cloud data, including the following steps, Classify and deploy the point cloud data format file to the corresponding terminal device according to the data resolution level and terminal demand; Dynamically determine the point cloud blocks that need to be loaded according to the user view range; Select the corresponding resolution level according to the terminal device type; Load the required point cloud data file block by block, store it in memory, and dynamically unload the blocks that have been loaded but are outside the user view range to release memory resources, and output the real-time loaded block point cloud data.

7. The multi-terminal oriented point cloud data format standardization conversion method of claim 6, wherein: According to the real-time loaded block point cloud data, dynamically adjust the resolution and block range of the point cloud data, and load new plugins to adapt to new point cloud formats, and output the optimized point cloud data, including the following steps, Capture the real-time operation of the user on the multi-terminal device and update the interaction state; Dynamically adjust the point cloud resolution level according to the user operation and viewpoint distance; According to the resolution level and view range, retrieve the new block data from the index table, and dynamically load the plugin of the new point cloud format through the modular plugin mechanism, and output the optimized point cloud data.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the multi-terminal-oriented point cloud data format standardization conversion method in any one of claims 1-7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the multi-terminal-oriented point cloud data format standardization conversion method in any one of claims 1-7.

Images