山东省寿光市滨海地区盐田信息提取方法研究

山东省寿光市滨海地区盐田水体因含盐度高,其光谱特征与海域水体及其他地物差异大,光谱特征显著；盐田系人为建造,排列整齐,几何特征明显,遥感影像上表现为纹理特征显著(棋盘状纹理、条纹状纹理),纹理指标可计算性强。首先采用缨帽变换方法增强光谱信息,采用定向滤波及灰度共生矩阵方法增强纹理信息;其次基于增强的光谱与纹理信息,采用以面向应用为目的的感兴趣地物提取方法对研究区TM图像进行分类,将分类结果与仅依据纯光谱及仅依据纯纹理分类结果相对比,分类总体精度分别为90.8985%、84.9102%和60.4017%。结果表明:以面向应用为目的的感兴趣地物提取方法分类精度最高。

Information Extraction Method of the Salt Field in the Coastal Areas in Shouguang, Shandong

Because of the high salinity,the spectral features of the salt field in the coastal areas in Shouguang of Shandong are greatly different from marine water bodies and other ground objects spectral features; the salt field is man-made,so,it has regular arrangement and significant geometric characteristics.In the remote sensing images,the above characteristics are shown as significant texture features (chessboard-like texture and stripe-like texture) and the texture indices have an excellent computability.In this paper,Firstly,using the tasseled cap transform to enhance spectral features of the salt field and using the directional convolution and the gray level co-occurrence matrix to enhance the texture features of the salt field.Secondly,based on the enhanced spectral and texture features,the TM image of the study area is classified by using the application-oriented interested ground objects extraction.The classification result is compared with the classification results based on spectrum only or texture only,the overall accuracy is respectively 90.8985%,84.9102% and 60.4017%.The result indicates that the accuracy of the classification method proposed in this paper is the highest.