基于水体透明度反演的太湖水生植被遥感信息提取

在使用多光谱遥感图像提取太湖水生植被分布时,由于水体中悬浮物和藻类等物质的影响,容易产生"异物同谱"现象,大面积水体被误分为沉水植被.本文首先通过TM图像反演太湖水体的透明度,基于RVI和NDVI植被指数,分别建立两类决策树,即透明度辅助的分类决策树和无透明度参与的分类决策树,将太湖分为水体、以浮叶植被为主导的水生植被和以沉水植被为主导的水生植被等三种类型.透明度辅助下的 NDVl 分类决策树方法,较好地消除分类过程中的"异物同谱"现象,是进行太湖水生植被分类的一种最好方法选择,把这种方法应用于2002年7月15日的Landsat ETM卫星遥感影像,结果表明太湖中以沉水植被为主导的水生植被约407.6km2,以浮叶植被为主导的水生植被约82.2km2.

Remote sensing information extraction of hydrophytes based on the retrieval of water transparency in Lake Taihu, China

There exists a phenomenon of different objects with a similar spectrum in the multispectral satellite images in Taihu Lake, which is caused by an influence of suspended matter and algae.So,it is easy to classify the water into submerged vegetations area by mistake.In this report,we firstly retrieved water transparency from Landsat TM image acquired on 15 July 2002.And then,the following two types of decision trees were developed on the basis of the two vegetation index of RVI and NDVI:(a)a transparency-assi...