CN117745910B - A fast rendering method for heat map of 3D model with massive grid points - Google Patents

Abstract

The invention discloses a method for quickly rendering a thermodynamic diagram of a three-dimensional model of a mass grid point in a blocking way, which comprises the steps of data file preprocessing, data file reading and thermodynamic value matching calculation, and the rendering speed of the thermodynamic diagram of the three-dimensional model of the mass grid point is improved on the basis of ensuring the rendering effect by means of spatial cluster analysis, file fragmentation, multithreading, strategy caching, blocking loading and the like. The data preprocessing stage is used for fragmenting the model grid point file and the point thermodynamic value file; in the data file reading stage, multithreading is adopted to read two types of file fragments, a spatial index tree STRtree is built based on point thermal value file fragments, a model grid block ModelBlock is built based on model grid point file fragments, and a built result is cached according to a certain caching strategy; and in the thermodynamic value matching calculation stage, based on the established STRtree, the thermodynamic values of the vertexes of each block ModelBlock grid are matched in parallel, the vertex colors are calculated, and the vertex colors of each block ModelBlock are returned to the client to realize the block thermodynamic diagram rendering.

Description

A fast rendering method for heat map of 3D model with massive grid points

Technical Field

The invention relates to a method and system for fast block rendering of a heat map of a massive grid point three-dimensional model, belonging to the technical field of three-dimensional space data visualization.

Background Art

Heatmap is a data visualization technology that uses color differences to show the size of values and can intuitively represent the concentration trend and distribution range of data points. In three-dimensional space, heatmaps are usually used to display data related to geographic location, such as temperature, pressure, or any other quantifiable indicators, and finally form a three-dimensional heat distribution map. This graphical representation method can help users quickly identify key areas of data. Heatmaps are particularly important in the application of digital twins of water conservancy projects. As a powerful tool, it can be used to monitor and analyze multi-dimensional information such as temperature distribution, water flow velocity, and pollutant concentration of water bodies in real time. This not only helps the daily management of water conservancy projects, but also provides real-time critical information support when responding to emergencies such as floods or water quality emergencies.

In the digital twin scenario of water conservancy projects, the complexity of physical entities leads to a huge number of grid points in their corresponding virtual three-dimensional models. These physical entities may include water networks, pumping stations, flood control systems, reservoirs and other water conservancy facilities, which need to be modeled with high precision due to their topographical or structural complexity. In order to accurately simulate and analyze the behavior and performance of these complex systems, it is necessary to create a detailed three-dimensional model containing millions or even tens of millions of grid points. Each grid point in the three-dimensional model represents specific data at that location, such as water flow velocity, water quality parameters, temperature, etc. These data need to be accurately represented and dynamically updated over time to reflect changes in actual conditions. Therefore, it is necessary to complete the rendering of thermal maps of three-dimensional models with massive grid points in a short period of time to support real-time analysis and decision-making in the digital twin of water conservancy projects.

Due to the complexity of physical entities, there are many difficulties in rendering heat maps for 3D models with massive grid points. On the one hand, the high density of grid points requires huge computing resources to process, and also requires the rendering system to be able to manage and maintain a large number of data streams. On the other hand, heat map rendering needs to convert the data of these grid points into a visual representation that users can understand. This process involves complex data-to-graphic mapping and the use of color coding. In addition, in order to meet the needs of practical applications, it is necessary to increase the rendering speed of heat maps for 3D models with massive grid points as much as possible, thereby reducing waiting delays and improving user experience. All these requirements make it a challenging task to render heat maps for 3D models with massive grid points in the digital twin scenario of water conservancy projects.

Summary of the invention

Purpose of the invention: To address the problem that rendering of heat maps of three-dimensional models with massive grid points consumes large resources and takes a long time to render, making it difficult to apply in actual engineering projects, a method and system for fast rendering of heat maps of three-dimensional models with massive grid points in blocks is provided. By combining various technical means, the overall rendering time is reduced while ensuring memory consumption and rendering effects, thereby improving user experience and providing key technical support for the application of digital twins of water conservancy projects.

Technical solution: A method for fast rendering of heat maps of 3D models with massive grid points in blocks. The method includes three stages: data file preprocessing, data file reading, and heat value matching calculation. It combines spatial clustering analysis, file segmentation, multi-threaded processing, strategy caching, block calculation and rendering, etc., to reduce the overall rendering time while ensuring memory consumption and rendering effect. The method includes the following steps:

Step 1) In the data file preprocessing stage, spatial clustering analysis is performed on the points in the point thermal value file to compress the number of points with similar thermal values in the same area, generate sparse point thermal value data, and reduce the amount of data to be processed;

Step 2) In the data file preprocessing stage, according to the principle of the same amount of shard data, the sparse point thermal value data is stored as multiple file shards to ensure that the amount of data processed by different threads is the same, so as to minimize the overall processing time;

Step 3) In the data file preprocessing stage, the model grid point file is first divided into multiple part slices according to the model parts to ensure the subsequent thermal map rendering effect; secondly, based on the principle of the same amount of data for approximate slices, each part slice is stored as multiple file slices of similar size;

Step 4) In the data file reading stage, multiple threads are used to read the point thermal value file slices, and multiple threads collaborate to build a spatial index tree STRtree during the reading process to reduce processing time;

Step 5) In the data file reading stage, use multi-threading to read the model grid point file slices and construct multiple model blocks ModelBlock;

Step 6) During the data file reading phase, the constructed STRtree and ModelBlock are cached according to a certain cache strategy to improve subsequent access speed while ensuring memory consumption;

Step 7) In the thermal value matching calculation stage, use multithreading to process the corresponding ModelBlock, match the coordinate values of the ModelBlock and the STRtree to calculate the thermal value of the mesh vertex and map the mesh vertex color;

Step 8) If a thread is processed, the grid color of the corresponding ModelBlock is returned to the client for rendering to achieve a block rendering effect.

Furthermore, the specific steps of performing spatial cluster analysis on the points of the point thermal value file in step 1) are as follows:

The point thermal value file is represented as the following set:

F _heat = {p ₁ , p ₂ ,..., p _N }

Where N is the number of rows of data in the point thermal value file, p _i =＜ _xi , _yi , z _i , υ _i ＞ is the i-th row of point data in the point thermal value file, x _i , _yi , z _i are the three-dimensional coordinate values, υ _i is the corresponding thermal value, and For the cth class in the kth round of clustering, the initial classification is defined as Then calculate the Euclidean distance matrix between the categories in the current clustering round:

Among them, _Nk is the number of classes in the kth round of clustering, d _a,b is the Euclidean distance between the ath class and the bth class, expressed as the Euclidean distance between the class centers of the ath class and the bth class, and the class center is expressed as the mean of the data within the class:

After calculating the distance matrix, assuming is the minimum distance, then and The two categories are merged and the next round of clustering operation is performed; after the clustering is completed, some point data in each category are randomly removed, and all the remaining point data are merged to obtain sparse point thermal value data

Furthermore, the specific steps of slicing the sparse point thermal value data in step 2) are as follows:

Sparse point thermal value data The number of data items included is The number of data entries contained in the _shard data is:

Where Num _split is the number of shards, then the kth shard data It is expressed as:

[k ^* Size _split : (k+1) ^* Size _split ] is a slicing operation, which means intercepting the data at the position from the k*Size _split index to the (k+1)*Size _split index, and after obtaining the slicing data, storing the slicing data as a slicing file.

Furthermore, the specific construction process of STRtree in step 4) is as follows:

First, initialize the set S = {s ₁ , s ₂ , ..., s _N }, where each element in S is a spatial object containing only one thermal value point data p _i . For the initialization set S, s _i .mbt = p _i .mbr, mbt is the minimum bounding cube of the spatial object;

Construct STRtree, set the threshold m of STRtree, which represents the maximum number of child nodes contained in a non-leaf node, and then divide the set S into γ = [N/m] spatial objects. If γ = 1, set S is used as the root node, and each element in S is used as a child node of the root node, and the construction process ends; if γ > 1, perform the spatial object division operation;

In the spatial object division, It represents the number of divisions in each direction. The division operation is as follows: sort each element _si in S in ascending order in the x direction and divide it into ε parts evenly; for each division result in the x direction, perform the above division operation in the y direction; finally, for each division result in the y direction, perform the above division operation in the z direction to complete the spatial object division and continue to build the STRtree.

Furthermore, in step 6), the process of caching the constructed STRtree and ModelBlock according to a certain caching strategy is as follows:

The LRU cache strategy is used to cache the STRtree and ModelBlock construction results. First, the cache queue Q _l is defined, where l is the cache capacity, and the STRtree and ModelBlock construction results are collectively referred to as elements e;

Before constructing e, first check whether element e exists in Q _l . If it exists, directly obtain and use it, skip the construction process of element e, and put element e at the head of queue Q _l to indicate the most recent access; otherwise, execute the construction process of element e;

After constructing element e, if the queue Q _l is not full, element e is inserted at the head of the queue; if the queue Q _l is full, the tail element is removed and element e is inserted at the head of the queue.

Furthermore, the specific steps of matching the coordinate values of ModelBlock and STRtree in step 7) are as follows:

Assume that ModelBlock contains m 3D mesh vertex information, that is, M = {g ₁ , g ₂ , ..., g _m }, where g _j = <x _j , y _j , z _j > is the mesh vertex coordinate. First, calculate the minimum bounding cube of ModelBlock:

M.mbr={min(X), max(X), min(Y), max(Y), min(Z), max(Z)}

Where X, Y and Z are the x, y and z coordinate sets of ModelBlock respectively. Traverse the STRtree nodes and match the minimum bounding cube of ModelBlock with the minimum bounding cube of the node. If the two intersect, place the thermal value point data included in the spatial object in the node into the result set R = {p ₁ , p ₂ , ..., p _n }. Match each 3D mesh vertex in ModelBlock with the result set R, and update the 3D mesh vertex g _j = <g _j .x, g _j .y, g _j .z, p _o .υ>, that is, match the thermal value for it, where p _o ∈ R satisfies:

Furthermore, the specific steps of mapping the mesh vertex colors in step 7) are as follows:

Define the maximum and minimum values of the thermal values of all mesh vertices in ModelBlock as v _max and v _min respectively. Use blue to correspond to v _min and red to correspond to v _max . For g _j ∈ M, define The mesh vertex color mapping formula is as follows:

red _j = [min(255, max(255*ratio _j ))]

green _j = [min(255, max(0, 255-abs(255*(ratio _j -0.5)*2)))]

blu% = [min(0, max(0, 255*(1-ratio _j )))]

Among them, red _j , green _j and blue _j are the red, green and blue component values of the mesh vertex g _j respectively.

A block-based rapid rendering system for heat maps of a massive grid point three-dimensional model includes a server for matching and calculating the heat values of the model grid points and a client for rendering the heat map based on the colors of the grid vertices.

In the data file preprocessing stage, the server uses spatial clustering analysis and equal-size sharding strategy to generate point thermal value file shards, and uses position sharding strategy and approximate-size sharding strategy to generate model grid point file shards;

In the data file reading stage, multi-threading is used to read the point thermal value file shards, and the spatial index tree STRtree is built during the reading process; multi-threading is used to read the model grid point file shards and build multiple model blocks ModelBlock; after the construction is completed, STRtree and ModelBlock are cached;

In the thermal value matching calculation phase, multiple threads match the coordinate values of the ModelBlock they built with the STRtree, set the grid point thermal value for each ModelBlock, and map the grid color based on the grid point thermal value. After the matching calculation of each ModelBlock is completed, the grid vertex color is returned to the client;

Each time the client obtains the vertex color of the ModelBlock mesh, it loads the heat map of the corresponding part based on the 3D engine to achieve block rendering of the 3D heat map.

The specific implementation process and method of the system are the same and will not be repeated here.

A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for fast block rendering of a heat map of a three-dimensional model with a large number of grid points as described above is implemented.

A computer-readable storage medium stores a computer program for executing the above-mentioned method for fast rendering of a heat map of a three-dimensional model with a large number of grid points.

Beneficial effect: Compared with the prior art, the present invention provides a method and system for fast block rendering of heat maps of three-dimensional models with massive grid points, and divides the rendering process of heat maps of three-dimensional models with massive grid points into three stages: data file preprocessing, data file reading, and thermal value matching calculation. Combining a variety of technical means, the rendering of heat maps of three-dimensional models with massive grid points is accelerated under the premise of ensuring memory consumption and rendering effect, and the user waiting delay is reduced, providing a technical basis for fast block rendering of heat maps of three-dimensional models with massive grid points. In the data file preprocessing stage, the thermal value file and the model grid point file are sliced and divided to provide a data basis for subsequent multi-threaded processing; in the data file reading stage, multi-threading is used to read the sliced file divided in the previous step, and STRtree and ModelBlock are constructed during the reading process, and the construction results are cached to ensure memory consumption while improving the overall processing speed and subsequent access efficiency; in the thermal value matching calculation stage, multi-threading is used to match ModelBlock with STRtree, the thermal value of the grid point is calculated and mapped to the grid point color, and the ModelBlock grid point color is returned to the client multiple times to realize block rendering. The present invention solves the problems of difficulty in rendering three-dimensional model heat maps of massive grid points, slow rendering speed, and large resource consumption, and provides a technical basis for rapid rendering of three-dimensional heat maps of massive grid points in blocks for digital twin water conservancy projects, which can better serve related industries.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a diagram of a method architecture according to an embodiment of the present invention;

FIG. 2 is a flow chart of a method according to an embodiment of the present invention.

DETAILED DESCRIPTION

The present invention is further explained below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

A fast block rendering method for heat maps of three-dimensional models with massive grid points, through spatial clustering analysis, file sharding, multi-threaded processing, strategy caching, block calculation and rendering, etc., can improve the rendering effect of heat maps of three-dimensional models with massive grid points while ensuring the rendering effect and memory consumption, providing basic technical support for digital twins of water conservancy projects.

As shown in FIG1 , the method for fast rendering of a heat map of a three-dimensional model with massive grid points is divided into three stages: data file preprocessing, data file reading, and thermal value matching calculation. The method for fast rendering of a heat map of a three-dimensional model with massive grid points described in FIG2 includes the following steps:

Step 1) In the data file preprocessing stage, spatial clustering analysis is performed on the point thermal value files, and the point thermal value files are represented as the following set:

F _heat = {p ₁ , p ₂ ,..., p _N }

Where N is the number of rows of data in the point thermal value file, p _i =＜ _xi , _yi , _zi , _υi > is the i-th row of point data in the point thermal value file, _xi , _yi , _zi are the three-dimensional coordinate values, _υi is the corresponding thermal value, let For the cth class in the kth round of clustering, the initial classification is defined as Then calculate the Euclidean distance matrix between the categories in the current clustering round:

Among them, _Nk is the number of classes in the kth round of clustering, d _a,b is the Euclidean distance between the ath class and the bth class, expressed as the Euclidean distance between the class centers of the ath class and the bth class, and the class center is expressed as the mean of the data within the class:

After calculating the distance matrix, assuming is the minimum distance, then and The two categories are merged and the next round of clustering operation is performed; after the clustering is completed, some point data in each category are randomly removed, and all the remaining point data are merged to obtain sparse point thermal value data

Step 2) In the data file preprocessing stage, according to the principle of the same amount of shard data, the sparse point thermal value data is sharded. The number of data items included is The number of data entries contained in the _shard data is:

Where Num _split is the number of shards, then the kth shard data It is expressed as:

Among them, [k*Size _split ：(k+1)*Size _split ] is a slicing operation, which means intercepting the data from the k*Size _split index to the (k+1)*Size _split index position, and after obtaining the shard data, storing the shard data as a shard file;

Step 3) In the data file preprocessing stage, the model grid point file is first divided into multiple part slices according to the model part to ensure the subsequent thermal map rendering effect; secondly, based on the principle of the same amount of data in the approximate slices, each part slice is divided into model slices of similar size and stored as multiple files to ensure that the amount of data processed by different threads is approximately the same, so as to minimize the overall processing time; Step 1)-Step 3) is an offline process and does not include the time consumption of thermal map rendering;

Step 4) In the data file reading stage, use multithreading to read the point thermal value file slices generated in step 1) and step 2), and collaboratively build a thread-safe STRtree during the reading process. First, initialize the set S = {s ₁ , s ₂ , ..., s _N }, each element in S is a spatial object containing only one thermal value point data p _i. For the initialization set S, s _i .mbr = p _i .mbr, mbr is the minimum bounding cube of the spatial object;

Construct STRtree, set the threshold m of STRtree, which represents the maximum number of child nodes contained in a non-leaf node, and then divide the set S into If γ = 1, the set S is taken as the root node, and each element in S is taken as the child node of the root node, and the construction process ends; if γ > 1, the spatial object division operation is performed;

In the spatial object division, Indicates the number of divisions in each direction. Sort each element _si in S in ascending order in the x direction and divide it into ε parts evenly. For each division result in the x direction, perform the above division operation in the y direction. Finally, for each division result in the y direction, perform the above division operation in the z direction to complete the spatial object division and continue to build the STRtree.

Step 5) In the data file reading stage, use multi-threading to read the model grid point file slices generated in step 3) to construct multiple ModelBlocks, and step 4) and step 5) are performed in parallel to improve the processing speed;

Step 6) In the data file reading phase, the LRU cache strategy is used to cache the STRtree and ModelBlock construction results. First, a cache queue Q _l is defined, where l is the cache capacity, and the STRtree and ModelBlock construction results are collectively referred to as elements e;

Before constructing e, first check whether element e exists in Q _l . If it exists, directly obtain and use it, skip the construction process of element e, and put element e at the head of queue Q _l to indicate the most recent access; otherwise, execute the construction process of element e;

After constructing element e, if queue Q _l is not full, element e is inserted at the head of the queue; if queue Q _l is full, the tail element is removed and element e is inserted at the head of the queue;

Step 7) Each thread in step 5) continues to process the corresponding ModelBlock and matches its coordinate value with the STRtree. Assuming that ModelBlock contains m three-dimensional mesh vertex information, that is, M = {g ₁ , g ₂ , ..., g _m }, where g _j = <x _j , y _j , z _j > is the mesh vertex coordinate, first calculate the minimum bounding cube of ModelBlock:

M.mbr={min(X), max(X), min(Y), max(Y), min(Z), max(Z)}

Where X, Y and Z are the x, y and z coordinate sets of ModelBlock respectively. Traverse the STRtree nodes and match the minimum bounding cube of ModelBlock with the minimum bounding cube of the node. If the two intersect, place the thermal value point data included in the spatial object in the node into the result set R = {p ₁ , p ₂ , ..., p _n }. Match each 3D mesh vertex in ModelBlock with the result set R, and update the 3D mesh vertex g _j = <g _j .x, g _j .y, g _j .z, p _o .υ>, that is, match the thermal value for it, where p _o ∈ R satisfies:

Define the maximum and minimum values of the thermal values of all mesh vertices in ModelBlock as vmax and vmin respectively. Use blue to correspond to _vmin and red to correspond to _vmax . For _gj∈M , define The mesh vertex color mapping formula is as follows:

Among them, red _j , green _j and blue _j are the red, green and blue component values of the mesh vertex g _j respectively;

Step 8) If a thread in step 7) is processed, the grid color of the corresponding ModelBlock is returned to the client for rendering to achieve a block rendering effect.

A method system for fast rendering of heat maps of three-dimensional models with massive grid points in blocks includes a server for matching and calculating heat values of model grid points and a client for rendering heat maps based on grid vertex colors.

In the data file preprocessing stage, the server uses spatial clustering analysis and equal-size sharding strategy to generate point thermal value file shards, and uses position sharding strategy and approximate-size sharding strategy to generate model grid point file shards;

In the data file reading stage, multi-threading is used to read the point thermal value file shards, and a thread-safe STRtree is constructed during the reading process; multi-threading is used to read the model grid point file shards and construct multiple model blocks ModelBlock; after the construction is completed, the STRtree and ModelBlock construction results are cached;

In the thermal value matching calculation phase, multithreading matches the coordinate values of the constructed ModelBlock with the STRtree, sets the grid point thermal value for each ModelBlock, and maps the grid color based on the grid point thermal value. After the matching calculation of each ModelBlock is completed, the grid vertex color is returned to the client;

Each time the client obtains the vertex color of the ModelBlock mesh, it loads the heat map of the corresponding part based on the 3D engine to achieve block rendering of the 3D heat map.

Obviously, those skilled in the art should understand that the various steps of the method for rapid rendering of a heat map of a three-dimensional model with a large number of grid points or the various modules of the system for rapid rendering of a heat map of a three-dimensional model with a large number of grid points in the above-mentioned embodiment of the present invention can be implemented by a general computing device, which can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented by executable program codes of computing devices, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules respectively, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the embodiments of the present invention are not limited to any specific combination of hardware and software.

Claims (6)

1. A method for fast rendering of heat maps of three-dimensional models with massive grid points by blocks, characterized in that the method includes three stages: data file preprocessing, data file reading, and thermal value matching calculation. A variety of means are used to improve the rendering speed of heat maps of three-dimensional models with massive grid points while ensuring the rendering effect of the heat map. The method includes the following steps: Step 1) In the data file preprocessing stage, spatial clustering analysis is performed on the points in the point thermal value file to compress the number of points with similar thermal values in the same area and generate sparse point thermal value data; Step 2) In the data file preprocessing stage, according to the principle of equal data volume of the shards, the sparse point thermal value data is stored as multiple file shards; Step 3) In the data file preprocessing stage, the model grid point file is first divided into multiple part slices according to the model parts; secondly, based on the principle of the same amount of data in the approximate slices, each part slice is stored as multiple file slices; Step 4) In the data file reading stage, multiple threads are used to read the point thermal value file slices, and multiple threads collaborate to build a spatial index tree STRtree during the reading process; Step 5) In the data file reading stage, use multi-threading to read the model grid point file slices in step 3) to construct multiple model blocks ModelBlock; Step 6) During the data file reading phase, the constructed STRtree and ModelBlock are cached; Step 7) In the thermal value matching calculation stage, use multithreading to process the corresponding ModelBlock, match the coordinate values of the ModelBlock and the STRtree to calculate the thermal value of the mesh vertex and map the mesh vertex color; Step 8) If a thread in step 7) is processed, the grid color of the corresponding ModelBlock is returned to the client for rendering to achieve a block rendering effect; The specific construction process of STRtree in step 4) is as follows: First, initialize the set S = {s ₁ , s ₂ , ..., s _N }, each element in S is a spatial object containing only one thermal value point data p _i , for the initialization set S, s _i .mbr = p _i .mbr, mbr is the minimum bounding cube of the spatial object; Construct STRtree, set the threshold m of STRtree, which represents the maximum number of child nodes contained in a non-leaf node, and then divide the set S into If γ = 1, the set S is taken as the root node, and each element in S is taken as the child node of the root node, and the construction process ends; if γ > 1, the spatial object division operation is performed; In the spatial object division, Indicates the number of divisions in each direction. Sort each element _si in S in ascending order in the x direction and divide it into ε parts evenly. For each division result in the x direction, perform the above division operation in the y direction. Finally, for each division result in the y direction, perform the above division operation in the z direction to complete the spatial object division and continue to build the STRtree. Step 6) In the data file reading phase, the LRU cache strategy is used to cache the STRtree and ModelBlock construction results. First, a cache queue Q _l is defined, where l is the cache capacity, and the STRtree and ModelBlock construction results are collectively referred to as elements e; Before constructing e, first check whether element e exists in Q _l . If it exists, directly obtain and use it, skip the construction process of element e, and put element e at the head of queue Q _l to indicate the most recent access; otherwise, execute the construction process of element e; After constructing element e, if queue Q _l is not full, element e is inserted at the head of the queue; if queue Q _l is full, the tail element is removed and element e is inserted at the head of the queue; Step 7) Each thread in step 5) continues to process the corresponding ModelBlock and matches its coordinate value with the STRtree. Assume that ModelBlock contains m three-dimensional mesh vertex information, that is, N = {g ₁ , g ₂ , ..., g _m }, where g _j = <x _j , y _j , z _j > is the mesh vertex coordinate. First, calculate the minimum bounding cube of ModelBlock, traverse the STRtree nodes, and match the minimum bounding cube of ModelBlock with the minimum bounding cube of the node. If the two intersect, place the thermal value point data included in the spatial object in the node into the result set R = {p ₁ , p ₂ , ..., p _n }, match each three-dimensional mesh vertex in ModelBlock with the result set R, and update the three-dimensional mesh vertex g _j = <g _j .x, g _j .y, g _j .z, p _o .v>, that is, match its thermal value, where p _o ∈ R satisfies: Define the maximum and minimum values of the thermal values of all mesh vertices in ModelBlock as v _max and v _min respectively. Use blue to correspond to v _min and red to correspond to v _max . For g _j ∈ M, define The mesh vertex color mapping formula is as follows: Among them, red _j , green _j and blue _j are the red, green and blue component values of the mesh vertex g _j respectively. 2. The method for fast rendering of heat maps of three-dimensional models of massive grid points according to claim 1 is characterized in that the specific steps of performing spatial clustering analysis on the points of the point heat value file in step 1) are as follows: The point thermal value file is represented as the following set: F _heat ={p ₁ , p ₂ ,…,p _v } Where N is the number of rows of data in the point thermal value file, p _i = < _xi , _yi , _zi , _vi > is the i-th row of point data in the point thermal value file, x _i , _yi , _zi are the three-dimensional coordinate values, and vi _is the corresponding thermal value. For the cth class in the kth round of clustering, the initial classification is defined as Then calculate the Euclidean distance matrix between the categories in the current clustering round: Among them, _Nk is the number of classes in the kth round of clustering, d _a,b is the Euclidean distance between the ath class and the bth class, expressed as the Euclidean distance between the class centers of the ath class and the bth class, and the class center is expressed as the mean of the data within the class: After calculating the distance matrix, assuming is the minimum distance, then and The two categories are merged and the next round of clustering operation is performed; after the clustering is completed, some point data in each category are randomly removed, and all the remaining point data are merged to obtain sparse point thermal value data 3. The method for fast rendering of a heat map of a massive grid point three-dimensional model according to claim 1 is characterized in that the specific steps of performing slicing processing on the sparse point heat value data in step 2) are as follows: Sparse point thermal value data The number of data items included is The number of data entries contained in the _shard data is: Where Num _split is the number of shards, then the kth shard data It is expressed as: [k*Size _split ：(k+1)*Size _split ] is a slicing operation, which means intercepting the data at the position from the k*Size _split index to the (k+1)*Size _split index, and after obtaining the slicing data, storing the slicing data as a slicing file. 4. A system for fast rendering of heat maps of three-dimensional models with a large number of grid points by blocks for implementing the method described in claim 1, characterized in that it includes a server for matching and calculating the heat values of model grid points and a client for rendering heat maps based on grid vertex colors; In the data file preprocessing stage, the server uses spatial clustering analysis and equal-size sharding strategy to generate point thermal value file shards, and uses position sharding strategy and approximate-size sharding strategy to generate model grid point file shards; In the data file reading stage, multi-threading is used to read the point thermal value file shards, and the spatial index tree STRtree is built during the reading process; multi-threading is used to read the model grid point file shards and build multiple model blocks ModelBlock; after the construction is completed, the STRtree and ModelBlock construction results are cached; In the thermal value matching calculation phase, multithreading matches the coordinate values of the constructed ModelBlock with the STRtree, sets the grid point thermal value for each ModelBlock, and maps the grid color based on the grid point thermal value. After the matching calculation of each ModelBlock is completed, the grid vertex color is returned to the client; Each time the client obtains the vertex color of the ModelBlock mesh, it loads the heat map of the corresponding part based on the 3D engine to achieve block rendering of the 3D heat map. 5. A computer device, characterized in that: the computer device includes a memory, a processor and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, a method for rapid block rendering of a heat map of a three-dimensional model of a massive grid point as described in any one of claims 1-3 is implemented. 6. A computer-readable storage medium having a computer program/instruction stored thereon, wherein the computer program/instruction, when executed by a processor, implements the steps of the method according to any one of claims 1 to 3.