高分辨率影像面向对象分类特征选择方法研究-地图制图学与地理信息工程专业论文.docxVIP

Subject ： Feature Selection for Object-oriented Classification of High Resolution Remote Sensing Images Specialty ：Cartography and Geographic Information Engineering Name ：YangNing （Signature） Instructor：Jia Jianhua （Signature） ABSTRACT Object-oriented image classification approaches, mainly including the nearest neighbor(NN) algorithm and membership function based on fuzzy, is a suitable for high spatial resolution image information extraction technology. It is chiefly characterized by small sample and high dimensionality of features. And the membership function o f fuzzy classification is more suitable for the larger image area. However, in the actual classification of membership function, according to the features of remote sensing images and experience knowledge, a small amount of features by analysis and try were used to classification. Finally, the classification result was good or not to determine whether the features would be selected. Obviously, it would cause subjectivity and blindness. And it isnt scientific and practical. Therefore, selected features could not ensure the classification speed and accuracy. This research was founded by ―Research on the land use type quickly extraction with high resolution data‖ (No: E0202/1112/0104) in the ―12th Five Year Plan‖. For the feature selection of object-oriented classification using membership function, this research mainly develops the following aspect to study: It was proposed that the feature selection algorithm to remove Irrelevant features. Based on small sample size, the Relief algorithms which is considered one of the best way in the feature selection area, were improved from aspects of drawing sample randomly, KNN denoising, iteration times in this paper. And 25 dimensions feature were selected from initial 63 dimensions feature based on the improved Relief algorithm during the experiment in the paper. The results showed that the improved Relief algorithm is more useful for feature selection in the