Researches On Change Detection From High Resolution Remote Sensing Images

The remote sensing image change detection is a technique to acquire change information of observed objects through contrast and analysis of remote sensing images of the same region at different time. With the development of the high resolution sensor technologies, the remote sensing image change detection gets widespread application. Change detection from remote sensing images will play a very important role in military applications such as the attack effect evaluation, monitoring of military targets, troops disposition and so on.Aiming at problems of shadow and projection difference elimination in change detection, the thesis proposes algorithms and methods by taking account of characteristics of high resolution remote sensing images. The effectiveness of the proposed methods are tested and verified by experiments.Characteristics of high resolution remote sensing images and their influence on change detection are summarized. A typical procedure of change detection and techniques involved in the procedure are overviewed. An experimental investigation into the performance of the change detection method based on texture features is conducted, showing that texture features are better than grayscale differences for change detection in reducing the influence of image registration errors. Effects of different texture features and texture windows to the performance of changed detection are also tested and analyzed in the experiment.An automatic algorithm to extract shadows of buildings from high resolution remote sensing images of urban area is proposed, tested and compared with the Ostu method and the empirical threshold two-valued method, showing that the new algorithm can improve the correctness ratio of detection and lower the false-alarm ratio, by keeping the shape and dimensions of a shadow well.A change detection method regarding shadow and projection difference is designed by integrating texture feature differences, automatic shadow extraction and projection difference handling. Experiments show good effects on the removal of the shadow influence on change detection based on texture features with reference to a shadow difference image. Advantages to elimination of projection difference influence on change detection are also shown in the experiments with a combination morphology method, which is easier to implementation than the method using change rate of a region.