Remote Sensing Image Fusion And Application Research

Remote sensing technique was proposed in the 1960s. Based upon the electromagnetic theory, this technique detects and recognizes objects by the electromagnetic wave which was reflected by the objects in long distance. With the help of the imagery produced by remote sensing technique, we can plot maps much faster, search targets, geodetic survey and so on, these made the data from remote sensing technique very widely used recently.Nowadays, in order to improve the precision of data and application results, the fusion method study is becoming a hot point. The fusion method means to combine existing technologies, create much more powerful tools. The remote fusion method means to fuse the data selected from all the existing sensors, the remote fusion method is normally divided into three levels: pixel level, feature level and the decision level.The goal of this paper is to research fusion algorithms, and detect objects by fusion method. The ship detection, as said in this paper, which is widely used in commerce, fishing, sea traffic and military monitor has great significance to the development of a broad prospect. Nowadays, single source ship detection has achieved great success, as for the SAR ship detection plays an important role in the SAR technology. But the single source detection is constrained by the ability of the sensor itself, just like the optical sensor is weak in penetrability, and the SAR imagery usually has a serious noise problem. This paper analyzed and improved the Dempster-Shafer Theory and realized the ship detection algorithm with the fusion result of SAR and optical imagery on feature level.The main contributions of this dissertation are:1.Realize typical fusion algorithms, such as the wavelet fusion, the HPFS fusion;2.Realize single source object detection;3.Realize ship detection system based on image fusion and improve the DS Theory.