| In the extraction of remote sensing-image information, the images are often only interested in certain parts. In order to analyze these goals of the images, they will be separated from the images, and on this basis it will be possible to further the goals of measurement and use of the image, that is, remote sensing-image segmentation technology which is one of the hot issues in the past 40 years.This article focuses on two algorithms of segmentation and their application in remote sensing-images: The adaptive threshold segmentation algorithm based on the wavelet transform and segmentation algorithm based on the level set. This article improves C-V algorithm of the level set ,and designs the demo system of remote sensing-image segmentation. This article's main contents are as follows:First, the remote sensing-image segmentation method is presented. This article summarizes the existing remote sensing-image segmentation method from four major aspects(based on the threshold segmentation method, based on edge detection methods, based on the regional segmentation method, based on specific theory and tool methods);Second, a new method of the application on remote sensing-image segmentation, that is, the adaptive threshold segmentation algorithm based on the wavelet transform is proposed. The article gives the theory to explain the algorithm in the image segmentation results and efficiency, and proves the algorithm is more obviously effective and practical.Third, the C-V method of level set which uses joint algorithm with FCM and C-V method is improved, and the experiment result shows that the improved algorithm not only eliminates the possibility that the original C-V algorithm can not effectively carry out the surface evolution, but also greatly reduces the segmentation time spent; Four, the demo system of remote sensing-image segmentation is designed. Based on actual needs of projects, improved algorithm is realized by the system which uses object-oriented manner. |
PDF Full Text Request