Research On 3D Terrain Rapid Reconstruction Technology Of UAV Remote Sensing Data

UAV remote sensing has Characteristics of low cost, simple operation, flexible and high-resolution images can be obtained,which develops rapidly in the area of Earth observation remote sensing applications. Since the complementary of these images in time and space information, using the relevant theorey and technic of computing visual and photographic measurements can be realized the 3D terrain reconstruction based on UAV remote sensing images.The paper analyzes the 3D terrain reconstruction of UAV remote sensing images, and obtains preliminary results. The main content is divided into the following several parts:1. The paper uses Graphic Processing Unit(GPU) to solve the time-consuming problem of feature points detection of UAV remote sensing images, thus realizes that the SIFT algorithm based on GPU is applied in UAV remote sensing image processing, and reduces the feature point detection time.2. To solve the problem of discrete point cloud model is not conducive to subsequent terrain analysis and visualization problem, the paper use local planar projection algorithm to realize the three-dimensional mesh reconstruction of UAV remote sensing terrain and use the GPU to accelerate processing algorithms to improve the reconstruction speed.3. Missing data of point cloud model of UAV lead to the mesh model reconstructed having holes which directly affect the subsequent display. To solve the problem of the holes in the grid, the paper using an improved algorithm for edge patching holes based on edge constraint, which achieve a large area of simple hole filling and improve the mesh quality...