Terrian Enhance Expression By Fusing The Ground-level Image And DEM

There are many aspects of the terrain expression in the real world of the current GIS that is hard to be accepted by ordinary people. A vertical view or an abstract map is often observed by the user of the current terrain expression. As far as the vertical view, the experience of people to observe the world is limited. Such view mode that separates the researchers’studies from the reality experience of ordinal people was once questioned because of the vision of God (pointed out by Longley).Meanwhile, such Geographical spatial cognition theory and expression is obviously contrary to the human cognitive habits.As far as the abstract expression, the expression of Map could be seen as an abstraction of the real world, which means there would be semantic differences between the real world and the abstract one. Meanwhile, the information abstracted by cartographers can also be lost and distorted, which would leads to some difficulties for the users in reading on the abstract map and concluding the geographical meanings. Thus, the organic integration of scientific observation and the expression which conforms to human habits would be the basis of GIS’s socialization.In this paper, we fuse the ground-level images with DEM by curve matching to implement the mapping between the2D image and the3D terrain surface. Thus, we implement the expression of the DEM area by high resolution photos and visualization of GIS data through the ground-level photos.Firstly, this paper designs a classification algorithm based on decision tree to extract skyline by splitting the sky and the ground from the ground-level image, and extracts terrain contour lines by image abstraction algorithm based on the classification; In this paper, we extract DEM skyline and contour line based the360°scene depth map of a sampled view point by using the ray tracing algorithmSecondly, the space parameter calculation algorithm of image is studied in this paper:in preliminary stage we matches photo skyline and DEM skyline to get the initial matching results by using the curve curvature integral descriptor; in the precise matching stage we matches photo contour lines and DEM contour lines to get the best matching position by cross correlation; the we derives the space parameters of photo by the parameter calculation formula;Finally, we can determine the photo in3d space position and orientation according to its space parameters. We build the mutual mapping relationship between photo pixels with three-dimensional points by Ray tracing algorithm to implement the photo mapping to3d terrain surface、GIS data shown in the photo and the measurement function in the photo. In this article, we supplement the traditional terrain expression through this new form of expression, which is more close to the perspective of ordinary people and would be impetus for the socialization of GIS.