基于决策树的浙江省河流水系遥感信息提取研究

Study on Remote Sensing Information Extraction of River System in Zhejiang Province Based on Decision Tree

现有的基于水体指数的遥感水体信息提取方法存在漏提和误提问题。以浙江省某流域为例,基于改进的水体指数法(WNDWI)和改进的谱间关系法,提出一种新的决策树水体提取模型,使用2015年四景Landsat/TM遥感影像,对该河流水系进行提取和分布制图。结果表明,与已有的水体指数提取方法相比,本方法可以更好地提取细小水体,减少阴影,提取精度更高。

There are problems in the current remote sensing water information extraction methods based on water body index.Through taking a river basin in Zhejiang Province as an example,the study proposes a new decision tree water extraction model on improved water body index method(WNDWI)and improved spectral relation method.Then the study extracts and maps the river system distribution with Landsat/TM remote sensing images of four landscapes in 2015.The results show that compared with the existing water index extraction methods,we can extract fine water better,and reduce shadow with higher extraction precision with this method.