Desertification Feature Extraction Based On Fuzzy C-means Algorithm Over Genetic Algorithm And Kernel Principal Component Analysis

At present, remote sensing technology has become a regional desertification monitoring and evaluation of the important means of regional desertification, soil, vegetation the main access to information means. Although remote sensing technology in the land desertification monitoring and analysis of changes have great advantages and become the study of desertification one of the indispensable means. However, to realize desertification rapid and accurate analysis and evaluation of the changes is not an easy task. Faced with such massive source of information, and how timely, accurate access to the necessary information and to use, how to explore new technology and more accurate method of remote sensing images from more access to the necessary information has been a topic of remote sensing image processing by the urgent need The key technology.For these purposes, the paper for traditional desertification in the extraction of information, often fail to use remote sensing data in high-end statistics, which overlook the number of pixels correlation between the non-linear, resulting in non-linear image information extraction areas The lack of access at home and abroad in a wide range of remote sensing technology, remote sensing image processing methods, desertification information extraction methods on the basis of literature, focusing on how to use modern mathematical methods to solve non-linear information from remote sensing images in the study. Make full use of existing data and information in the previous desertification information extraction on the basis of research results, respectively, the fuzzy clustering analysis, genetic algorithm and the principal component of the theoretical methods such as research, for the above method has the advantage with less than , The combination of remote sensing data in the characteristics of high-end statistics, based on the design GMFCM (based on the Mahalanobis distance fuzzy c-means clustering genetic algorithm) and KPCA (kernel principal component analysis), feature extraction model.Will be based on GMFCM (based on the Mahalanobis distance fuzzy c-means clustering genetic algorithm) and KPCA (kernel principal component analysis), feature extraction model were applied to the Inner Mongolia Autonomous Region in Yikezhaomeng - Alashanmeng- Bayanzhuoer: the western part of Kubuqi Desert and north-east of Ulanbuhe Desert's 1989 and 2000 TM images ETM images feature extraction, rather than the traditional feature extraction method, the better the extraction of non-linear features, feature extraction and image results Over the past 12 years the actual local desertification changes in comparison anastomosis, verify the validity of the model.Example results show that the results of research on enriching the feature extraction of desertification, desertification for the feature extraction method provides a new and effective ways possible.