基于穗光谱指数的水稻产量预测

Rice yield prediction with panicle spectral indices

为探索运用水稻穗光谱植被指数预测水稻产量的可行性,以2个水稻品种为材料,设置3个氮素水平,测定了3个时期水稻叶片和穗的高光谱反射(350~2 500 nm)和色素含量,并测定了水稻的产量构成组分和籽粒产量。结果表明:与典型的植物反射光谱相比,水稻穗的反射光谱具有"绿峰消失"的特征;与叶片光谱指数[归一化差值指数(normalized vegetation index,NDVI)和光化学反射指数(photochemical reflectance index,PRI)]相比,穗光谱指数对叶绿素更敏感,而且能更准确地区分氮素水平。水稻叶片NDVI和PRI预测产量的均方根误差(RSME)分别为873.4~1 125.0、723.3~889.4 kg·hm^(-2),而穗NDVI和PRI预测产量的RSME分别为681.7~743.1、515.0~637.8 kg·hm^(-2),表明水稻穗光谱指数比叶片光谱指数更适合于水稻产量预测。

This pilot study was aimed to investigate the feasibility of predicting rice yield with panicle spectral indices. A field experiment was conducted with two rice genotypes of contrasting yield potential and three contrasting nitrogen( N) levels at Hangzhou,China in 2015. Leaf and panicle hyperspectral reflectance( 350-2 500 nm),chlorophyll concentrations( Chlc) and carotenoids concentrations( Carc) were measured at three different dates,and the yield components and grain yield were determined at maturing stage. It was found that the panicle spectra were distinctly different from the leaf spectra with the disappearance of the sharp green "hump". As compared to the corresponding leaf indices,the panicle normalized vegetation index( NDVI) and photochemical reflectance index( PRI)were more reliable in differentiating the N levels,more sensitive to the chlorophyll content,and performed more accurately in predicting the rice yield. In the rice yield prediction based on best relationship fit,NDVI and PRI in the leaves at the three dates yielded a RSME of 873. 4-1 125. 0 and 723. 3-889. 4 kg·hm^(-2),respectively,while NDVI and PRI in the panicles at the three dates achieved a RMSE of 681. 7-743. 1 and 515. 0-637. 8 kg·hm^(-2),respectively.