【摘要】 马尾松毛虫是我国南方地区最主要的森林害虫,给生态、经济与社会带来极大威胁。随着遥感数据国产化步伐的不断加快及空间、时间、光谱分辨率的不断提升,马尾松毛虫害空间识别技术的突破迎来新的契机。旨在探索一种适于南方山地丘陵区马尾松毛虫虫害信息提取的方法,其主要思路是:在实现马尾松林信息提取的基础上,选择并获取与马尾松毛虫虫害信息相关的地形因子、NDVI、叶面积指数及红边参数等指标,基于光谱实现松毛虫信息片层的分割;以光谱片层为对象,进行主成分变换,提取其纹理特征,并利用决策树分类规则实现马尾松毛虫虫害信息的提取。更多还原

【Abstract】 Dendrolimus punctatus is the main forest pest in the southern hilly areas in China,which brings about severe threat to the ecology,economy and society.With the accelerating pace of China’s localization of remote sensing data and the rising of spatial,time,and spectral resolusions,the spatial recognition techonology’s breakthroughs for Dendrolimus punctatus damage are facing a new opportunity.This paper aims to explore a method for Dendrolimus punctatus damage information extraction suitable for southern hilly areas in China.The main ideas are: the terrian factors,NDVI,leaf area index and red edge parameter which are relevant to the Dendrolimus punctatus damage information were selected and extracted on the basis of the extration of masson pine forest;then the film subset was achieved based on the spectral features by taking the spectra film as the objects and conducting the principal component analysis.These process will realize the Dendrolimus punctatus damage information extraction relying on the Decision Tree Classification.更多还原