论文部分内容阅读
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。
Dendrolimus punctatus is the most important forest pest in southern China, which brings great threats to ecology, economy and society. As the localization of remote sensing data continues to accelerate and space, time and spectral resolution continue to increase, breakthroughs in space recognition technology for Masson pine caterpillars will usher in new opportunities. Aiming at exploring a method suitable for extracting information of Dendrolimus punctatus in mountainous hilly area of southern China, the main idea is: based on the information extraction of masson pine, selecting and acquiring the topographical factors related to pest information of Dendrolimus punctatus, NDVI , Leaf area index and red-edge parameter, and then the information layer of the pine caterpillars was divided based on the spectrum. The spectral components were taken as the object to transform the principal components, and the texture features were extracted. Information extraction.