基于高光谱微分指数的杉木炭疽病病情指数反演  

Inversion Anthracnose Disease Indices of Chinese Fir Based on Hyperspectral Derivative Indices

作　　者：伍南[1] 刘君昂[1] 周国英[1] 闫瑞坤[1] 董文统[1] 

机构地区：[1]中南林业科技大学林业生物技术湖南省重点实验室,长沙410004

出　　处：《林业科学》2012年第8期94-98,共5页Scientia Silvae Sinicae

基　　金：林业公益性行业科研专项经费项目(200904006;201004014)

摘　　要：通过实地调查,获取杉木炭疽病病情指数以及各级病情梯度的高光谱遥感数据,将冠层光谱一阶微分数据与相应的病情指数进行相关性分析,选取相关系数大于0.7的高光谱微分指数构建反演模型,并对模型进行精度检验。结果表明:病情指数与冠层光谱一阶微分值在431～485nm,549～616nm和700～754nm3个波段达到极显著相关;以微分指数SDy,SDr,SDg/SDb,(SDg-SDb)/(SDg+SDb)和(SDr-SDy)/(SDr+SDy)为变量反演病情指数的最佳模型均为线性模型,均方根差在0.094～0.119之间。该研究结果表明,利用高光谱微分指数估测杉木炭疽病病情指数精度较高,且具有较高的实际应用价值。The anthracnose disease index （DI） of Chinese fir and the corresponding hyper-spectral remote sensing data of the disease different severity were collected through the field investigation. The correlation between DI and the first derivative data of the canopy spectral was analyzed. Inversion models were built by selecting correlation coefficients higher than 0.7, and their precision was tested. The result showed that there was high correlation between DI and the first derivative data in the regions of 431 -485 nm, 549 -616 nm and 700 -754 nm; and the linear model contained SDy, SDr, SDg/SDb, （SDg - SDb）/（SDg ＋ SDb） and SDr- SDy）/（SDr ＋ SDy） was the best one, with the RMS （Croot mean square） error ranged from 0. 094 to 0. 119. This study showed hyper-spectral derivative data could estimate the disease severity of Chinese fir with a high the precision. This study has a high practical application value.

关 键 词：高光谱 微分指数 杉木炭疽病 病情指数 反演模型 

分 类 号：O433.4[机械工程—光学工程] S763[理学—光学]