土地覆盖遥感专题信息的分层提取方法及其应用

The Multi-layer Information Extraction of Landcover from Remotely Sensed Imagery

重点讨论了土地覆盖遥感专题信息的分层提取方法，即根据各类目标的不同特征，采取相应的信息提取方法，分别建立专题信息层，最后把各专题层合并得到整体分类图；在此方案实施过程中，对各类目标开展了野外遥感调查和特征系统分析，并对样本区进行了ＧＰＳ定位，以获取空间辅助信息，提高提取精度；分层提取方法充分考虑到各类目标的不同特点，从而避免了通常分类方法中。

The paper discusses the multi layer information extracting method for landcover from remotely sensed imagery. According to the features of different targets, multi layers of thematic information are acquired by using corresponding ways. Then the landcover classification map is completed by overlaying them. During the process, we carried out field investigation and analyzed the target features,and positioned the sample areas by GPS, so as to obtain the ancillary information and improve the classification result. This method fully considers the target features, so that the mixture of different types in the map is avoided significantly.