Research On A GPU-based Parallel Implementation Of Image Resampling And An Optimization Study Of Tile Cache Strategy

With the improvement of remote sensing technology,the data volume of remote sensing images is growing at a geometric progression rate.Remote sensing image data has gradually become the main source of spatial data updating.At the same time,with the maturation of network technology,sharing remote sensing image data through the network has become a reality.The area of the share of remote sensing image data has also shifted from professional applications to the public life.However,currently the share of remote sensing image data faces problems such as the organization of the spatial data and the restriction of the network bandwidth.Therefore,researching on the organization of spatial data and the prefetching cache technology in network transmission is the key to achieving efficient sharing of remote sensing image data.Based on the analysis of the organization of spatial data and the prefetching cache technology in network transmission,this thesis studies a kind of data structure based on the tile and hierarchy structure and the prefetching cache for spatial data.The main contents are as follows:(1)Efficient spatial data organization and management—Application of the tile and hierarchy structure in remote sensing image data processingThis thesis investigates the image texture and establishes a region-based image interpolation method.Based on the theory of the image threshold segmentation,this method divides the image into two regions:the smooth region and the complex region and different image interpolation algorithms are used in different regions.Then,with parallelism nature of data processing for remote sensing image data,this thesis takes the advantage of the parallelism and the programmability of GPU to improve the calculation speed of the spatial interpolation.Finally,an approach of data subarea based on linear quadtree coding is realized in this thesis.(2)Fast transmission of remote sensing image data on the internet—Design the network GIS integrated with web prefetching and cachingThis thesis investigates the prefetch techniques of tiles data.Using the powerful calculation ability of a single computer,web browsers can prefetch tiles data to the client so as to shorten the response time.Then,using the spatial relation of tiles data,this thesis establishes a cache strategy which is based on the heat of time-frequency and geographic location.This strategy separates the heat of time-frequency from the heat of geographic location as two parts and studies the multi-attribute decision-making method to manage web server's buffer space.Finally,this thesis integrates prefetching and caching in the network GIS and realizes high-speed transmission of remote sensing image data on the internet.(3)An actual application system development—Map service subsystem of the Rural Land Contract Management Information System of Zhejiang ProvinceAccording to the application requirements,this thesis studies the organization and the network transmission of spatial data,which was applied to the Rural Land Contract Management Information System of Zhejiang Province.The result indicates the study in this thesis is feasible.