Research On Geographical Raster Data Compression And Geographical Scene Management

As huge amount of geographical data especially 3D terrain data being produced, it has become a serious problem to visualize large-scale geographical scene. Nowadays, large-scale geographical scene visualization mainly acts as a platform for visualizing infrastructural geographical information, which faces many problems such as transmission of huge amount of data, large in-memory scene structure, resulting in non-enough efficiency for visualization. Especially when a great deal of geo-environmental elements and 3D models loaded, reasonable frame rate requirement is hard to achieve.The frame rate in large-scale geographical scene visualization mainly depends on data transmission and in-memory scene structure. Data compression can reduce the data to be transmitted, and a cache strategy helps avoid repeated data request. According to traditional LOD method, much scene data can be removed from memory, but neither in time nor efficient enough without considering the adaptivity for different LODs.In this article, data compression, transmission and realtime visualization of large scale geographical scene are studied. The contributions and relevant work in the paper are described as follows.Firstly, an improved image compression method combining CABAC based on SPIHT is presented. More contexts from SPIHT are used to reduce redundant data. When applied to DEM compression, storage and transmission efficiency will be improved.Secondy, a data compression and management framework for visualization is adopted, which is managed by TMS format based on data pyramid structure. After being compressed, data to be stored and transmitted is greatly reduced, hence it will improve the visualization performance especially in low bandwidth network.Thirdly, an adaptive scene graph management method based on LOD is proposed. According to the distance from eye to scene, different scene graph size is adaptively configured. At the same time, a two-level cachy strategy is used in scene visualization. All cache data are managed by memory cache manager, and swap between memory and disk.Finaly, All the methods adopted are integrated in osgEarth platform, and are proved to be efficient.