基于HJ-1CCD遥感影像的西双版纳橡胶种植区提取

摘要

The rubber plantation area in Xishuangbanna was extracted by the supervised classification method from the HJ remote sensing images in the 1st leaf-umbrella of rubber tree color to stable stage with specific spectral features.This method solved the spectral differences problem of supervised classification caused by age,species and farming.The non-extractive features ware removed before supervised classification which effectively reduced the number of categories and workload.Analysis showed that the classification accuracy was up to 97.6％ which reflected preferably rubber cultivation distribution in Xishuangbanna.It was a good model for the promotion and application of domestic satellite.%利用橡胶林第一蓬叶变色-稳定期环境遥感影像作为遥感信息源,根据橡胶林这一时期特有的影像特征,通过监督分类方法实现对西双版纳橡胶种植分布区信息的提取,解决了监督分类时树龄、品种和耕作制度造成的光谱差异问题;监督分类前利用NDVI阈值对非提取目标进行剔除,有效减少了分类数和工作量.分析表明,本次提取精度达到97.6％,较好地反映了西双版纳橡胶种植分布情况,同时对国产卫星的推广应用具有良好的示范作用.