基于QuickBird遥感影像的棚户区提取与制图

摘要

This paper used GLCM method that maps informal settlement areas according to their texture information by QuickBird remotely sensed imagery in Xiaguan District,Nanjing,China.First,the selected four texture features（contrast,energy,homogeneity and correlation） were computed within a local neighborhood to construct a feature space.Second,they were then fed to an unsupervised classifier（ISODATA algorithm） with a fixed number of output classes.Third,the morphological operation was performed to divide the joint small areas and adjust fringe region of the informal settlement areas.In experiments for study area,the results demonstrated the effectiveness of the proposed method in extracting and mapping informal settlements.Moreover,the results of this paper could render planning support and monitor informal settlements for local authorities.%以QuickBird遥感影像为数据源,从纹理信息的角度,利用灰度共生矩阵（GLCM）的方法对南京市下关区的棚户区信息进行了提取。首先,利用所选择的4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调整和优化,从而获得研究区域的棚户区信息。研究结果表明,基于纹理特征的灰度共生矩阵方法对于棚户区的提取和制图是有效可行的,同时本文的研究结果可以为南京市棚户区的管理及城市规划提供更为科学的依据。