Organization Of Large Scale Terrain And Rectification And Mapping Of Remote Sensing Image Texture

The simulation training system, the war analysis and the tactical rules research system based on combined military operation of consolidative modern multi-branch of the armed forces, urgently needs to construct the large-scale three-dimensional virtual battlefield scene. The large-scale three-dimensional geography scene is the important foundation to construct the large-scale virtual battlefield scene. The spatial scope of large-scale virtual battlefield is very big, the three-dimensional terrain data quantity and the surface image texture data quantity extremely is all huge, however, in order to guarantee the large-scale three-dimensional geography scene has the good reality and to achive the real-time display and interactive roam of the large-scale three-dimensional geography scene, it causes the modelling and the rendering of the large-scale three-dimensional geography scene faced with very challenges.This paper, based on the large-scale three-dimensional terrain rendering algorithm and the software foundation in the topic-based group, according to the demand of constructing the large-scale virtual battlefield natural scene, studies the large-scale terrain data organization, the remote sensing image texture rectification and the high-resolution texture mapping thoroughly and so on the key technologies. The paper completes the prime task and the research results are as follows:1. This paper discusses and analyses two international terrain data formats, which are Digital Elevation Model (DEM) and Digital Terrain Elevation Data (DTED). Then we summarize a practicable algorithm that can translate DEM into DTED.It analyses MATLAB's interpolate function, completes the terrain data's interpolation using it's function and get good result.2. This paper has studied the remote sensing image texture and the three-dimensional terrain's data alignment question, designed and completed one kind solution of rectifying remote sensing image through selecting the ground control point to simplify solution system of linear equations coefficient which can product the rectification equation. This method compares the low accuracy the wide range remote sensing image texture to adjust suitably, in order to enable the texture image in the correct mapping terrain, first formulates the control point selection strategy, then selects the ground according to the strategy the control point to carry on the geometry fine adjustment again. The experimental result indicated that, this algorithm obtained the good remote sensing image texture rectification effect, has satisfied the reality request.3. This paper has studied the organization method of the large-scale three-dimensional terrain data. For exploiting the GPU's performance fully, enhancing the rendering efficiency, it organizes the terrain data with the hierarchical structure of terrain tile pyramid (TTP) and terrain summary pyramid (TSP). It completes and compares two kinds of algorithm to produce TTP on the ground of analyzing the TTP's characteristic expression and the TTP's store structure. It selects the good Bottom-Up method to produce the store structure of terrain tile pyramid (TTP) and the terrain summary pyramid (TSP). Although this method increased the preliminary data's processing time, but can obviously enhance the terrain rendering and the texture mapping efficiency.4. In order to strengthen the ground the third dimension and enhances the texture mapping's efficiency, the paper organized the high resolution remote sensing image texture with the texture tile quadtree's hierarchical structure, and presentated high resolution texture dispatch mapping algorithm based on the terrain summary pyramid (TSP). Through the TSP which is resident in memory it selects each rendering terrain block, finds corresponding texture block using the match algorithm, while the Hash table is responsible for buffer scheduling of the texture block And completes texture mapping through GPU's vertex shader. The experimental result indicated this algorithm can complete the wide range, the high resolution remote sensing image texture mapping highly effectively.Based on above research results, this article developed a sort of software of large-scale three-dimensional geography scene mapped with remote sensing image texture. It has good reality and interactivity.