基于MATLAB的遥感图像棚户区识别与提取

The Identification and Extraction of Informal Settlement Areas From Remote Sensing Images Based on MATLAB

以QuickBird遥感影像为数据源,研究在MATLAB环境下,利用灰度共生矩阵(GLCM)对QuickBird遥感图像棚户区纹理特征进行识别和提取的方法。首先,利用所选择的四个纹理特征统计量(对比度、能量、同质性和相关性)构建一个特征空间。然后运用非监督分类的方法(ISODATA算法)将研究区分成确定的类别数目。最后,根据实际情况,利用数学形态法对分类结果进行精确定位,从而提取得到研究区域的棚户区信息。

This paper studies the identification and extraction informal settlement areas from remote sensing images using ray level co-occurrence matrix（GLCM）,realized the classification and analysis for QuickBird remote sensing image using the filter of the un-supervised classification methods under MATLAB environment.Firstly,by using the selected four texture statistics（contrast,energy,homogeneity and relevance） to construct a feature space.Secondly,the unsupervised classification method（ISODATA algorithm） will research into determining the number of classes.Finally,according to the actual situation,using mathematical morphological method for the classification results for accurate positioning,so as to extract obtained from the study area shantytown information.