纹理主方向的遥感影像居民地提取  

Residential area extraction from remote sensing image based on texture principal directions

作　　者：金飞[1] 张占睦[1] 芮杰[1] 

机构地区：[1]信息工程大学测绘学院,郑州450052

出　　处：《测绘科学》2010年第4期139-141,共3页Science of Surveying and Mapping

摘　　要：本文研究利用基于傅立叶变换和Hough变换的灰度共生矩阵方法对遥感影像居民地进行提取。首先,通过傅立叶变换把遥感影像变换到频率域,然后利用Hough变换找出遥感影像的主纹理方向,进而把影像主纹理方向旋转至水平,最后利用灰度共生矩阵方法进行纹理分割。实验结果表明:基于傅立叶变换和Hough变换的灰度共生矩阵方法对遥感影像居民地有较好的提取结果。A remote sensing image residential area extracting method based on Fourier transformation and Hough transformation grey level eo-ocurrence matrix is presented in this paper. First, the RS image is translated to the frequency domain by Fourier transformation, then Hough transformation is used to extract the texture principal direction, so the texture principal direction of the image is rotated to horizental direction, at last the texture is segmented by grey level co-ocurrenee matrix method. Experimental resuhs show that good residential area extraction results can be acquired by our method.

关 键 词：傅立叶变换 HOUGH变换 灰度共生矩阵 纹理分割 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]