基于小波变换和SVMOLI图像分类方法.docVIP

基于小波变换和SVMOLI图像分类方法 　　摘 要 以陕西省靖边县2013年 8月份的Landsat 8 OLI图像作为基础数据源，比较最大似然法和本文提出的分类方法识别研究区10种主要地物类型的精度。首先对OLI图像进行像素级融合处理和3层分解的小波滤波；之后对图像进行LBV变换；依据地物间光谱特征和形态特征差异为10种地物类型选择样本、执行SVM监督分类和基于数学形态学开闭运算的分类后处理；选取总体分类精度和Kappa系数两项指标，将所得分类图与对最大似然法分类图比较。结果显示，本文提出的分类方法能够降低OLI图像分类后的椒盐效应，分类图的总体分类精度和Kappa系数分别为：82.75 %、0.773，较最大似然法分类结果分别提高：14.72%和16.42%。基于小波变换的SVM监督分类方法能够较为准确的识别研究区10种主要地物类型，抑制分类后的椒盐效应，适合用于在陕北黄土高原梁峁沟壑区解译OLI图像。 　　关键词 OLI图像；黄土高原梁峁沟壑区；椒盐效应；小波滤波；LBV变换 　　中图法分类号： TP753；P237.4 文献标识码： A 文章编号： 　　Classification of OLI Images based on Wavelet Transform and SVM 　　LIU Wei 　　（College of Information Engineering， ，Xizang Minzu University， Xianyang 712082， China） 　　Abstract：This research aims to seek out the most suitable method for classification of Landsat OLI multispectral remote sensing images acquired in August， 2013， by the comparison study of supervised classification based on maximum likelihood and the method proposed in this paper. There are 10 kinds of land use type in OLI images of Jingbian County. Firstly， OLI images were fused with panchromatic image and then processed by means of 3 levels wavelet filtering. Secondly， LBV transform was applied to OLI images. Thirdly， training samples set for each kind land use type were collected， and then supervised classification based on SVM ， opening-closing operation in mathematical morphology were carried out to get precise information of each kind land use type. Finally， assessing the classification results for method proposed in this paper and maximum likelihood， by overall classification accuracy and Kappa coefficient as evaluation indexes.Results show that： the overall accuracy and Kappa coefficient of classified image using the method proposed in this paper were 82.75% and 0.773， with growth of 14.72% and 16.42% compared with classification image using maximum likelihood. Meanwhile， removal of salt and pepper noise in classified images was more effective using method proposed in this paper. Higher classification accuracy