摘要
为从SPOT-5图像上提取苏州河网,设计了一种按层次进行的河网结构类型识别方法。首先,针对苏州河网总体结构特征,选择出典型判别因子;然后,用其中某一个或几个判别因子进行判别,按照统计特征分成可能的几个类别,再在结果类别中利用其它因子进行下一步判别。这样,以二分法的方式逐层深入地进行分类,最后达到河网结构类别识别的目的。
This paper has designed a structural type identification method of the Suzhou river network picked up from the SPOT-5 image. According to the overall structural features of the Suzhou river network, some typical discrimination factors are selected, and then discrimination is made by using one or several discrimination factors. On the basis of the statistical features, it can be divided into some possible classifications and, by using other factors, differentiation can be further carried out in the resulting classification. Therefore, with the dichotomy, the types of river network can be classified. This method proves to be very effective in the structural type identification of the Suzhou river network and can attain the aim of differentiating structures of river networks.
出处
《国土资源遥感》
CSCD
2004年第4期33-35,45,共4页
Remote Sensing for Land & Resources
基金
国家自然科学基金重点资助项目(项目编号:50239030)。
