Extraction Of Remote Sensing Image Planar Target Based On Mathematical Morphology

In recent years, with the rapid development of remote-sensing technology, it's possible for us to obtain abundant spatial information, but how to deal with and interpret such vast spatial information automatically or semi-automatically is a big problem in the informational course of the whole society. This article does research on semi-automatic extraction of some area ground features, such as water area, habitation, vegetation, etc. The main research work is concluded as follows:1. Expatiating the background and the development of the research, analyzing the basic technique of extracting area ground feature according to the object's characteristic;2. Introducing the conception, character, basic operation of two-value morphology, grey morphology and control two-value morphology , and then introduce the morphologic transformation formed by basic operation upwards, finally the application of morphology in image processing;3. Do research on edge detection algorithm that is based on " rug-covered method" , with comparing the result with other detection algorithms, a conclusion can be made that slope intercepting method has some advantages in edge detection;4. After analyzing the texture segmentation model that has been brought forward by former researchers, combining the texture feature of the habitation in remote-sensing image the author brings forward two texture segmentation models that are based on uniting mathematical morphology, wavelet pyramid image and grey-accreted matrix, and then precision evaluation is made;5. For texture of water area is relatively exquisite and its' grey is relatively uniform, the author brings forward method of using high-hat and low-hat in image segmentation and method of combining optimum threshold two-value grey chart and logical operation of close and open operation, and then analyses the advantages and disadvantages of the two methods;6. As vegetation texture is relatively rough and its' grey is relatively uniform, the author brings forward method of single-point extraction based on morphology gradient.