CN114549761A - Real-scene three-dimensional model layered rendering optimization method and system based on distributed storage and storage medium - Google Patents

Abstract

The invention discloses a layered rendering optimization method, a layered rendering optimization system and a layered rendering optimization storage medium for a live-action three-dimensional model based on distributed storage, wherein the method comprises the steps of firstly, acquiring live-action three-dimensional data; then, segmenting the live-action three-dimensional data to obtain a plurality of image layers; finally, calling a rendering engine to load and render each layer according to a preset mode; and loading all layers in the live-action three-dimensional data. According to the layered rendering optimization method and system for the live-action three-dimensional model and the storage medium, the live-action three-dimensional model is divided into a plurality of layers to be respectively rendered and loaded, so that the rapid rendering process of the live-action three-dimensional model is realized, the size of the layer divided by the live-action three-dimensional model is set according to the equipment requirement, the rendering process is accelerated, and the rendering efficiency is improved; the phenomenon that the three-dimensional model with large data volume is blocked in the rendering process is effectively prevented.

Description

Distributed storage-based layered rendering optimization method and system for 3D model of reality, and storage medium

technical field

The invention relates to the technical field of architectural three-dimensional modeling, in particular to a distributed storage-based layered rendering optimization method and system for a three-dimensional model of a real scene, and a storage medium.

Background technique

In urban construction, it is necessary to provide a large number of architectural design schemes. The architectural design scheme is generally designed by 3D software. The overall effect of the architectural design scheme needs to be quickly displayed during the review of the architectural design scheme. The 3D image data designed by the virtual 3D software The amount of 3D image data is very large, and the display of 3D image data not only requires the support of expensive hardware equipment, but due to the different configurations of the display equipment, different rendering engines are generally used in different computer equipment. Graphs, or the 3D reality map of the planning and design of a township, are all massive data. The existing 3D reality loading and rendering methods are not suitable for the loading process of these large volumes of data, and the rendering engine may take a long time to complete the rendering process. , and even the display device is stuck during the display process. These problems will seriously hinder the efficiency of the 3D model display. In order to quickly load the 3D model under the condition of using ordinary hardware devices, it is necessary to use a smooth Rendering optimization method for displaying 3D models.

SUMMARY OF THE INVENTION

In view of this, the purpose of the present invention is to provide a distributed storage-based real-scene 3D model layered rendering optimization method, system, and storage medium. Fast rendering process for realistic 3D models.

For achieving the above object, the present invention provides the following technical solutions:

The distributed storage-based layered rendering optimization method for a 3D model of a real scene provided by the present invention includes the following steps:

Obtain real 3D data;

Divide the real 3D data to obtain several layers;

Allocate each layer to each terminal device for storage according to the distributed storage method, and build a distributed database;

Set the layer loading method and call the rendering engine to load and render each layer according to the preset method;

Until all layers in the reality 3D data are loaded.

Further, the distributed storage of the real 3D data is formed according to the following steps:

Build a MongoDB database;

Cut the 3D model data into several small pieces of 3D image 3Dtiles;

Form 3Dtiles into hierarchical catalog files and store them in the MongoDB database in batches;

Build an indexed database for accessing or recalling data.

Further, the 3D real scene data is divided into several rectangular blocks according to the spatial area, and the size of the block can be set to 800-1200 3D tiles.

Further, the loading method is implemented using edge computing technology, and the specific steps are as follows:

Receive rendering instructions sent by the cloud platform;

Invoke the rendering engine to load the layer data stored in the local memory according to the rendering instruction;

Transfer the rendered image data to the cloud platform.

Further, each layer is allocated to each terminal device for storage in a distributed storage manner, as follows:

Calculate each layer to obtain the data set of each layer;

Obtain the hardware resources of distributed terminal equipment;

Determine the matching relationship with the load balance of hardware resources of each terminal according to the layer data set and hardware resources;

Store the segmented layers in the corresponding terminal device according to the matching relationship;

Build the layer data storage index database.

Further, the segmentation of the real scene 3D data is performed according to the following steps:

Build image sequences of different levels according to the 3D model data, set the size of each level of images, and build a pyramid structure;

Based on Tile technology, each level of image is cut and divided into blocks to form small image Tiles;

Based on the file storage system, establish a hierarchical directory storage structure to store each small image Tiles;

According to the given 3D model and the latitude and longitude coordinates of the location point, the name of the tile to which it belongs and the relative storage path of the tile are calculated, so as to realize the fast query and acquisition service of 3Dtiles data;

Using the architecture of a distributed storage system, a hierarchical directory is established to store and manage the small image Tiles.

Further, the segmentation of the three-dimensional data of the real scene is performed according to the type of the model library components; the components in the model library include any one of buildings, water systems, traffic, realms, terrain, landforms, vegetation, pipelines, fences, and independent objects. one or more combinations.

Further, the following steps are also included:

Set the bottom material interface, the bottom material interface is used to call the material properties of the rendering model, and modify the RGB values of different components in the model, or modify the light map.

The layered rendering optimization system for a 3D model of a real scene provided by the present invention includes a memory, a processor, and a computer program stored in the memory and running on the processor. the steps of any one of the methods.

The storage medium provided by the present invention stores a computer program thereon, and when the program is executed by a processor, implements the steps of the method described in any one of claims 1-8.

The beneficial effects of the present invention are:

The distributed storage-based layered rendering optimization method, system, and storage medium for a 3D model of a real scene provided by the present invention utilizes dividing the 3D model of the real scene into several layers for rendering and loading, so as to realize the rapid rendering process of the 3D model of the real scene. It is required to set the size of the layer divided by the 3D model of the real scene, which speeds up the rendering process and improves the rendering efficiency, and effectively prevents the 3D model with a large amount of data from being stuck in the rendering process.

Other advantages, objects, and features of the present invention will be set forth in the description that follows, and will be apparent to those skilled in the art based on a study of the following, to the extent that is taught in the practice of the present invention. The objectives and other advantages of the present invention may be realized and attained by the following description.

Description of drawings

In order to make the purpose, technical solutions and beneficial effects of the present invention clearer, the present invention provides the following drawings for description:

FIG. 1 is a flowchart of a method for optimizing layered rendering of a 3D model of a reality based on distributed storage.

FIG. 2 is a distribution diagram of the layer structure of the 3D model of the real scene stored in a distributed manner.

FIG. 3 is a flowchart of a method for constructing a distributed database.

FIG. 4 is a schematic diagram of cloud computing-based terminal device distribution rendering.

Detailed ways

The present invention is further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

As shown in FIG. 1 , the distributed storage-based layered rendering optimization method for a real 3D model provided by the present embodiment, the method for optimizing real-time rendering in a game engine based on real 3D, includes the following steps:

Obtain real 3D data;

Divide the real 3D data to obtain several layers;

Allocating each layer to each storage device according to distributed storage; constructing a distributed database for storing each layer index database, and forming a loading data stream according to the layer index database;

Set the loading method of the loaded layer and call the rendering engine to load and render the loading data stream according to the loading method;

Until all layers in the real 3D data are displayed.

The divided layers in this embodiment are distributed and stored according to the following steps:

Calculate each layer to obtain the data set of each layer;

Obtain the hardware resources of distributed terminal equipment;

Determine the matching relationship with the load balance of hardware resources of each terminal according to the layer data set and hardware resources;

Store the segmented layers in the corresponding terminal device according to the matching relationship;

Build the layer data storage index database.

The loading level of each layer is set when the real 3D model of this embodiment is divided, and the loading level includes a low-level of detail model block and a high-level of detail model block.

The layer loading method in this embodiment adopts LOD loading, as follows:

Determine the display viewing angle range in the display environment;

Determine the layer loading level within the viewing angle range and the layer loading level outside the viewing angle range; the loading level within the viewing angle range is a high level of detail model nugget, and the loading level outside the viewing angle range is a low level of detail model nugget ;

The loading level outside the viewing angle range first determines the type of the loaded layer, and then determines the type of the bottom level of detail model block according to the layer type.

The loading process of each layer in this embodiment may adopt a distributed computing loading process. First, a loading task is allocated to each terminal device through the cloud platform, and each terminal device performs a three-dimensional reconstruction of a local scene according to the received layer loading task, and summarizes them into The cloud platform merges 3D models to generate a complete 3D model.

In this embodiment, data transmission adopts a combination of central cloud and edge computing, and uses edge computing to share more core network traffic and computing power, and optimize data transmission and signal processing processes.

In this embodiment, the hardware resources of each terminal device are analyzed and the data sets are matched and processed, thereby solving the problem of hardware differences of different terminal devices; the matching relationship between the hardware resources of each terminal device and the storage layer data set is used. , to ensure the hardware load balance of each terminal device, and improve the cross-platform processing efficiency.

The hardware resources in this embodiment refer to GPU attributes of each terminal device.

In this embodiment, different terminal devices are used to perform corresponding layer rendering, and the large-scale real 3D data to be rendered is divided into layers for distributed rendering, and the rendering engines of different terminal devices are used for rendering respectively, so as to obtain faster visualization. Effect.

At the same time, through multi-level extraction of 3D model file data, the loading of different levels is determined according to the viewing area, which reduces the amount of rendering data, improves the loading speed when the viewing angle moves, reduces the phenomenon of dropped frames, and achieves the best display effect of ordinary PCs. and viewing experience.

In this embodiment, the real-scene 3D data is segmented in a semi-automatic manner, so that the load of each layer is balanced, and at the same time, the top-level merging of the real-scene 3D data is removed, that is, the overview image data of the image is deleted;

The rendering engine provided by this embodiment adopts the game engine UE4;

The preset mode of the layer provided by this embodiment adopts the concurrent loading of partitions to form a data stream;

As shown in FIG. 2 , the 3D real scene data in this embodiment is divided into several blocks according to the real scene model in different areas of the space, and the size of each block can be determined according to the performance of the device. The size of the block in this embodiment can be It is set to 800-1200 tiles, and the optimal size of the tiles in this embodiment can be set to no more than 1000 tiles; then, the tiles located in different areas of the model space are divided into different areas, and the data streams of different areas are executed concurrently for input to the rendering engine for loading.

In this embodiment, a splitting and merging-based segmentation algorithm is used to segment the real scene three-dimensional data, and specifically, an octree-based splitting and merging algorithm or an adaptive bounding box-based splitting and merging algorithm may be used.

In this embodiment, the three-dimensional data 3DTiles of the real scene are obtained through the three-dimensional data cutting algorithm; the 3DTiles are to divide the space into blocks, and each block is called a "tile", that is, a tile.

The three-dimensional real scene data provided by this embodiment adopts distributed storage, and the distributed storage adopts MongoDB for storage, and the three-dimensional real scene model of the distributed storage is formed according to the following steps:

Build a MongoDB database;

Acquire 3D model data and cut the 3D model data into several small 3D image 3Dtiles;

The file format of the 3Dtiles in this embodiment includes files of type .json; .b3dm;

Form 3Dtiles into hierarchical catalog files and store them in the MongoDB database in batches;

Build an index database for accessing or recalling data;

The index database constructed in this embodiment is used for invoking and accessing distributed data services;

The cutting of the three-dimensional model data provided by this embodiment is performed according to the following steps:

Image sequences of different levels are constructed according to the three-dimensional model data. In this embodiment, the levels of the image sequence can be set to N levels, where N is a natural number less than 18; each level corresponds to an image of a specific size, and the higher the level, the smaller the corresponding image. That is, the clearer the image is formed, the first level is the top layer, and the 18th level is the bottom layer; finally, a pyramid structure is formed.

Based on Tile technology, each level of image is cut and divided into blocks to form small image Tiles;

As shown in Figure 3, based on the file storage system, a hierarchical directory storage structure is established to store each small image Tiles;

According to the given 3D model and the latitude and longitude coordinates of the location point, the name of the tile to which it belongs and the relative storage path of the tile are calculated, so as to realize the fast query and acquisition service of 3Dtiles data;

Adopt the architecture of the distributed storage system, establish a hierarchical directory, and store and manage the small image Tiles;

The method provided by this embodiment also includes the following steps:

Set the rendering engine of the web browser, and display the small 3D images 3Dtiles called from the distributed database, namely the MongoDB database, in the web browser after fusion processing;

The rendering engine in this embodiment may be constructed by using a multi-engine fusion technology, the multi-engine includes a game engine and a 3D engine, the game engine adopts the game engine ue4, and the 3D engine adopts the WebGL framework and engine.

The multi-engine can be constructed by virtual packaging or physical packaging, so as to realize the display of the three-dimensional model of the building, determine different rendering engines according to the number of scene display surfaces, automatically replace the engine to display the three-dimensional scene, and use the feature loading and matching of different rendering engines. The matched 3D scene image quickly displays the 3D model.

The segmentation of the real 3D data provided in this embodiment can also be performed according to the types of components in the model library; the components in the model library include buildings, water systems, traffic, boundaries, terrain, landforms, vegetation, pipelines, fences, and independent features The components are decomposed into geometric features and texture features. The loading sequence can load the geometric features of various components first, and then load the texture features of various components; or divide the components into different types and load important components first. The geometric features of the type member are loaded with the geometric features of the remaining members, and finally the texture features of these members are loaded.

As shown in FIG. 4 , the rendering engine optimization method provided in this embodiment improves the loading speed of the real 3D data and realizes large-scale real 3D data by dividing the real 3D data to be rendered and loading them concurrently according to different standards. The release and browsing of real-world 3D data overcomes the technical difficulty of fast loading of huge amounts of real-world 3D data.

The rendering optimization method provided by this embodiment avoids the loading of the model schematic diagram at the beginning of loading, improves the loading speed, and sets the underlying material interface, and the underlying material interface is used to call the material properties of the rendering model. The interface can be based on actual Personalize the settings according to the situation, further optimize the rendering effect if necessary, and modify the RGB values of different components in the model, or modify the lightmap.

This method optimizes the memory in the rendering process, and a good rendering effect can be achieved by using 32G or 20G of memory.

The above-mentioned embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or transformations made by those skilled in the art on the basis of the present invention are all within the protection scope of the present invention. The protection scope of the present invention is subject to the claims.

Claims (10)

1. The layered rendering optimization method of the live-action three-dimensional model based on distributed storage is characterized by comprising the following steps: the method comprises the following steps:

acquiring live-action three-dimensional data;

segmenting the live-action three-dimensional data to obtain a plurality of layers;

distributing each layer to each terminal device for storage according to a distributed storage mode, and constructing a distributed database;

setting a layer loading mode and calling a rendering engine to load and render each layer according to a preset mode;

and loading all layers in the live-action three-dimensional data.

2. The method of claim 1, wherein: the live-action three-dimensional data distributed storage is formed according to the following steps:

constructing a MongoDB database;

cutting the three-dimensional model data into a plurality of small three-dimensional images 3 Dtiles;

forming a hierarchical directory file by using 3Dtiles and storing the hierarchical directory file into a MongoDB database in batches;

an index database for accessing or calling data is constructed.

3. The method for optimizing the layered rendering of the live-action three-dimensional model based on the distributed storage according to claim 1, wherein: the real three-dimensional data divides the real three-dimensional data into a plurality of rectangular image blocks according to the spatial region, and the size of the image blocks can be set to 800-.

4. The method for optimizing the layered rendering of the live-action three-dimensional model based on the distributed storage according to claim 1, wherein: the loading mode is realized by adopting an edge computing technology, and the specific steps are as follows:

receiving a rendering instruction sent by a cloud platform;

calling a rendering engine according to the rendering instruction to load the layer data stored in the local memory;

and transmitting the rendered image data to the cloud platform.

5. The method for optimizing the layered rendering of the live-action three-dimensional model based on the distributed storage according to claim 1, wherein: the method for allocating the layers to the terminal devices for storage according to the distributed storage mode specifically comprises the following steps:

calculating each layer to obtain a data set of each layer;

acquiring hardware resources of distributed terminal equipment;

determining a matching relation with each terminal hardware resource load balance according to the layer data set and the hardware resources;

storing the divided layers in corresponding terminal equipment according to the matching relation;

and establishing a layer data storage index database.

6. The method for optimizing the layered rendering of the live-action three-dimensional model based on the distributed storage according to claim 1, wherein: the real scene three-dimensional data is segmented according to the following steps:

constructing image sequences of different levels according to the three-dimensional model data, setting the size of each level of image, and constructing a pyramid structure;

cutting and blocking each level of image based on Tile technology to form small blocks of images Tiles;

establishing a layered directory storage structure to store the small images Tiles based on a file storage system;

calculating the name of the Tile and the relative storage path of the Tile according to the given three-dimensional model and the longitude and latitude coordinate values of the position points, and realizing the rapid query and acquisition service of the 3Dtiles data;

and establishing a hierarchical directory by adopting a distributed storage system architecture, and storing and managing the small images Tiles.

7. The method for optimizing the layered rendering of the live-action three-dimensional model based on the distributed storage according to claim 1, wherein: the real scene three-dimensional data is divided according to the type of the model library component; the members in the model library comprise any one or a plurality of combinations of buildings, water systems, traffic, boundaries, terrains, landforms, vegetation, pipelines, palisade grids and independent ground objects.

8. The method for optimizing the layered rendering of the live-action three-dimensional model based on the distributed storage according to claim 1, wherein: further comprising the steps of:

and setting a bottom layer material interface, wherein the bottom layer material interface is used for calling the rendering model material attribute, and modifying the RGB values of different components in the model or modifying the illumination map.

9. A realistic three-dimensional model layered rendering optimization system comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the program, when executed by the processor, performs the steps of the method of any one of claims 1 to 8.

10. Storage medium on which a computer program is stored which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 8.

Images