Self-Adaptive Replacement and Pre-dispatch Algorithm Considering Spatial Relationship Constraint for 3D Spatial Data

Abstract

Since processing of large scale 3D spatial data is far exceeding the current calculation capacity of computer hardware, it is impossible that the whole 3D data could be loaded into memory at one time during 3D scene display and analysis visualization. Dynamic data dispatch is needed according to the scope of real-time 3D scene. In this case, traditional cache processing methods are inefficient. Aiming at the feature of spatial data, a new cache management algorithm is proposed based on spatial relationship constraint. Its kernel idea is to build spatial index for objects in the cache considering their spatial position factors, and thus improve the traditional cache replacement algorithm which only based on time and hit rate. On the other hand, 3D spatial data pre-dispatch method considering influence factor is put forward by extending the cache spatial index with R-tree to spatial object sample tree. Integrated module with this cache replacement and pre-dispatch approach become a part of 3D spatial data engine as a database management tool. The practice in 3D urban planning information system of Wuhan city, central China, has proved the validity and efficiency of this method on dynamic dispatch and analysis processing with large scale spatial data with size of more than Tb.