基于知识的山东丘陵区土地利用/覆盖分类研究

土地利用/覆盖信息的获取是土地利用/覆盖变化研究的前提和基础, 传统的基于光谱信息的分类无法克服地物光谱特征相似造成的混淆。以龙口市为例, 探讨了综合应用高程、坡度等地学专家知识和地物的光谱知识, 对山东丘陵地区土地利用/覆盖进行自动分类的方法。实验证明, 基于知识的土地利用ö覆盖分类方法消除了单纯利用光谱信息的不足, 达到了90. 24% 的分类精度, 远高于最大似然法分类。

Study on Knowledge-based Land Use/Cover Classification of Shandong Hilly Region

Land use/cover information is the precondition and foundation of LUCC(land use/cover change) which is indispensable to globe change research.Traditional classification method such as spectral-based classification can not solve the confusion of spectral similar problem.The classification method based on knowledge may sufficiently use remote sensing expert thought and experiences to improve remote sensing image classification accuracy.This article took Longkou city as an example,using spectral characteristic of land use/cover object and combining altitude,slope information,a geographic expert knowledge based classification method of Shandong hilly region was discussed.The study showed that,this method eliminates the deficiency of using spectral information simply,and the classification accuracy reached to 90.24%,which is much higher than maximum likely classification and proved that the classification method based on knowledge is efficient.