A Study On The Fast Height Information Extraction Of The Man-made Feature Based On The Remote Sensing Image

This paper makes a study of the man-made features' height information obtaining methods based on the remote sensing image. The related theories and technical methods of the man-made features' height information obtaining methods based on the single remote sensing image and the stereoscopic image are studied systematically, and the emphasis is put on the vertical edge extraction method and the probability relaxation matching method constrained by the TIN based on the point feature. The main contents and innovations are as follows:1. This article makes a introduction of the research background and significance. The status quo and development tendency of the man-made features' height information obtaining methods are analyzed.2. The principal of the man-made features' height information obtaining method based on the projection error principium in photogrammetry is analyzed. The principal and method of the man-made features' height information obtaining method based on the collinearity equation is put forward. The feasibility of the methods is proved through the experiments.3. On the basis of analyzing the principle of conventional Canny operator algorithm, the adaptive edge-detection method based on the canny operator is proposed; The advanced OPTA thinning algorithm which improves the thinning effects is put forward. The man-made features' vertical edges are extracted with the advanced Hough transform method, the experiment results show that the above proposed methods are all valid.4. After analyzing the principle and the process of the scale invariant feature transform(SIFT) feature based matching method, a new adaptive relative orientation method is proposed with the help of related image matching methods. The DSM generation method which uses the probability relaxation matching method constrained by the TIN based on the point feature is put forward. Based on the DSM and the topographic map data, the man-made features' height information is obtained.