采用不同光谱波段宽度的归一化植被指数估算水稻叶面积指数

通过大田试验,测量了水稻不同生育期的冠层光谱,并通过在TM红光波段和近红外波段范围内不断扩展波段宽度,计算了各个波段宽度对应的归一化植被指数(NDVI)及其与叶面积指数(LAI)的最大相关系数,再根据NDVI与LAI最佳拟合方程的最大R2值随波段宽度的变化特征确定了最佳波段宽度.结果表明:NDVI与LAI的相关系数和最佳拟合方程的最大R2值具有相同的变化趋势,在小于红光波段宽度60nm范围内,均随波段宽度的增大而减小,即窄波段NDVI比宽波段NDVI对LAI的估算效果好;在超过60nm后,由于受近红外波段宽度的影响而略有波动.对R2规律的分析表明,使用NDVI估算水稻LAI的最佳波段宽度为15nm.

Estimation of rice LAI by using NDVI at different spectral bandwidths

The canopy hyperspectral reflectance data of rice at its different development stages were collected from field measurement,and the corresponding NDVIs as well as the correlation coefficients of NDVIs and LAI were computed at extending bandwidth of TM red and near-infrared(NIR)spectra.According to the variation characteristics of best fitted R2 with spectral bandwidth,the optimal bandwidth was determined.The results showed that the correlation coefficients of LAI and NDVI and the maximum R2 of the best fitted functions at different spectral bandwidths had the same variation trend,i.e.,decreased with increasing bandwidth when the bandwidth was less than 60 nm.However,when the bandwidth was beyond 60 nm,the maximum R2 somewhat fluctuated due to the effect of NIR.The analysis of R2 variation with bandwidth indicated that 15 nm was the optimal bandwidth for the estimation of rice LAI by using NDVI.