基于光谱特征的SPOT-5影像马尾松毛虫虫害信息提取

摘要

A study was conducted to extract the information of Dendrolimus punctatus damage to Pinus massomana in Shaxian County of Fujian Province. Based on spectral features, the spectral monitoring indexes, such as normalized difference vegetation index, ratio vegetation index, green vegetation index, soil-adjusted vegetation index and red edge parameter, were constructed for series model of insect damage and inversion with a SPOT-5 multi-spectral image. And then the insect pest information was derived, and the total precision was 70.75%. The results show that it is practicable to apply remote sensing images to damage monitoring of D. punctatus.%以福建沙县作为研究区,以单期SPOT-5多光谱影像为数据源,基于其光谱特征,构建归一化植被指数(NDvi)、比值植被指数(Rvi；)、绿度植被指数(Gvi)、土壤调节植被指数(Savi；)、红边参数等光谱指数作为光谱监测指标,通过建立马尾松毛虫虫情级数模型并进行反演,进而提取虫害信息,其虫害信息提取的总精度为70.75％.研究结果证明利用单时相遥感数据进行马尾松毛虫虫害监测是可行的.