面向对象的遥感影像植被信息提取技术研究

摘要

Based on elaborating the basic principle and method of object-oriented remote sensing classification method, the classification training and precision evaluation of vegetation in Hongshan Region, Wuhan City in 2006 were carried out. The shadow removal method was adopted in order to solve the problem of different object with the same spectrum, and the quantification determination method was used to extract vegetation in image. The results showed that, the vegetation information extraction precision by object-oriented multi-scale segmentation is obviously higher than that of non-supervision classification precision, which is an effective extraction technique.%在阐述面向对象的遥感分类方法的基本原理和方法的基础上,对2006年武汉市洪山区的遥感影像数据中的植被进行了分类训练和精度评价.为了解决植被中存在的异物同谱的现象采用了阴影消除的方法,同时采用定量测定的方法提取影像中的植被.结果表明,采用面向对象的多尺度分割的植被信息提取精度明显高于非监督分类的分类精度,是一种行之有效的分类提取技术.