A tile-based fast loading method for 3D model data

Abstract

The invention discloses a three-dimensional model data tile type rapid loading method. At first, a three-dimensional model is divided into tiles, the three-dimensional tiles are organized according to an octree structure, the index file and the tile file are respectively stored, concerning the organization manner and display characteristics of the three-dimensional tile data, and from the point of view of reducing the real-time tile request loading frequency and increasing the cache efficiency, the tile data loading and calling time can be reduced by means of cache displacement and a cache prediction algorithm, so that the tile data display is efficient and smooth. Aiming at the problem of low loading speed of large-scene three-dimensional data, a method of three-dimensional model data tile type rapid loading is proposed and plays an important role in the three-dimensional scene roaming speed, the three-dimensional scene scheduling smoothness, and man-machine interaction experience.

Description

technical field [0001] The invention belongs to the technical field of geographic space information systems, in particular to a tile-type fast loading method for three-dimensional model data. Background technique [0002] Three-dimensional GIS (Geographic Information System) is an important development direction of GIS at present. It can perform intuitive expression of multi-resolution, multi-scale and multi-time and space three-dimensional scenes. Driven by high-tech such as precise modeling and virtual reality, it has obtained Steady development. The establishment of 3D scenes with high-precision, multi-resolution, multi-time-space, and multi-level-of-detail models produces massive, complex and multi-scale 3D model data, which makes the loading and display of large-scale 3D scenes lagging and slow. The efficient loading of 3D scenes is one of the most basic and key technologies, which directly affects the development prospects of 3D technology applications. [0003] In o...

AI Technical Summary

Problems solved by technology

The three most direct replacement strategies are: the first one, the least recently used LRU (Least Recently Used), removes the tiles that have not been accessed for the longest time, but may keep the tiles that users occasionally access in the cache; the second The second type, the least frequent use of LFU (Least Frequently Used), removes the tiles with the least total number of visits, but may keep the tiles that are frequently accessed in the early stage but not frequently accessed in the cache for a long time; the third type, advanced FIFO (First In First Out) removes the first stored tile, but ignores a user's long-term and high-frequency access to a specific area, and may delete a certain area that the user is interested in

The neighbor selection markov chain (NSMC) uses the moving direction of the first k tiles as the transition state, assuming that all tiles have the same transition probability, the model is simple, and the storage overhead is small, but the spatial data organization form and spatial features are not fully considered. the importance of

Images