论文部分内容阅读
针对全极化SAR影像的建筑区特性,提出了一种基于极化特征共生矩阵的城区建筑密度分析方法。首先将极化特征与共生矩阵结合,在考虑建筑区极化散射机理和建筑朝向作用的同时,兼顾了建筑区的空间排列信息,在此基础上为了增强建筑密度的局部区域特性,将共生矩阵特征进行K-means聚类,结合图像分块形成标号直方图统计矢量,进而对该直方图统计矢量进行矢量量化实现SAR影像城区的建筑密度分级。RadarSat-2全极化SAR影像城区建筑密度分析的实验表明,该方法既适用于建筑朝向复杂城区也适用于建筑排列整齐城区的密度信息提取。
Aiming at the characteristics of the building area of the fully polarimetric SAR image, a method of urban building density analysis based on the polarization characteristic co-occurrence matrix is proposed. Firstly, the polarization characteristics are combined with the co-occurrence matrix. Considering the polarimetric scattering mechanism of building area and the direction of building orientation, the information of spatial arrangement of building area is taken into consideration. Based on this, in order to enhance the local regional characteristics of building density, Feature K-means clustering, combined with the image block to form a label histogram statistical vector, and then the vector of the histogram vector to quantify the SAR image of urban construction density grading. Experiments on density analysis of RadarSat-2 fully polarimetric SAR imagery show that this method is suitable for both density-oriented information extraction of buildings with complex urban areas and neatly arranged urban areas.