论文部分内容阅读
基于面向对象和区域增长的方法,开展东营市城乡建设用地提取。通过分析局部斑点统计特征计算真实的纹理影像,并对强度数据进行多尺度分割,结合纹理影像信息和局部后向散射强度信息,应用面向对象和区域生长方法提取城乡建设用地。利用Google Earth提供的影像进行对照验证,生成混淆矩阵来评价城乡建设用地提取结果。在东营市的高分辨率TerraSAR-X数据的实验结果验证了本方法的有效性,提取精度达92.89%。
Based on the methods of object-oriented and regional growth, the construction of urban and rural construction land in Dongying City was carried out. The real texture image is calculated by analyzing the statistics of local speckles, and the multi-scale segmentation of intensity data is combined with texture image information and local backscattering intensity information to extract urban and rural construction land using object-oriented and region growing methods. Use the images provided by Google Earth for verification and generate the confusion matrix to evaluate the extraction results of urban and rural construction land. The experimental results of the high resolution TerraSAR-X data in Dongying verify the validity of this method with an accuracy of 92.89%.