无损检测植物叶片水分对植物生理生化研究及灌溉管理和旱情监测等均具有重要意义。 利用Gaia Sorter近红外高光谱仪(900～1 700 nm), 以不同生育期的60个鲜活玉米叶片为试验材料, 对叶肉不同区域的平均光谱及烘干称重法得到的水分含量分别用偏最小二乘法(PLS)及逐步多元线性回归(SMLR)进行建模分析。 结果表明, 验证集决定系数/标准偏差分别为0.975/1.18和0.980/1.02, 均取得较好的预测效果, 可实现单个玉米叶片平均含水量的测定; SMLR优选的特征波长(1 406和1 692 nm)建模预测结果表明, 利用高通量近红外相机结合滤光片方法实现玉米叶片冠层或高空遥感测量的可行性。 同时, 进行了叶片不同区域水分含量的成像分析, 结果表明, 验证集中6个叶片的叶肉与主叶脉区域水分含量的参考均值和预测均值的相关系数均达到0.85以上, 预测结果与实际情况相符合。

Non-destructive detection of plant leaves water content is of significance to plant physiological, biochemical research, irrigation management and drought monitoring. In this paper, Gaia Sorter Near-Infrared Spectrometer (900~1 700 nm) was used to detect the water content of 60 fresh corn leaves in different growth stages using PLS and SMLR models. The results demonstrated that the R2/SEP of validation set were 0.975/1.18, 0.980/1.02, all achieving better predictive results, which could bring out the determination of a single corn leaf average water content. The results of the SMLR model established with the preferred characteristic wavelength (1 406 nm, 1 692 nm) indicated that the use of high-throughput near-infrared camera combining filter method achieved the feasibility of corn leaves canopy or high-altitude remote sensing measurement. Simultaneously, the imaging analysis of the water content in different regions of the leaves was carried out, and the results revealed that the correlation coefficients between the measured mean values and the predictive mean values of the mesophyll and the main vein of the six leaves were 0.85, and the predictive results were in accordance with the actual situation.