Estimation of plant water content in cut chrysanthemum using leaf-based hyperspectral reflectance

Abstract

Water plays an important role in the growth process of cut chrysanthemum (Chrysanthemum morifolium Ramat.). Accurate monitoring of plant water content (PWC) is a vital guarantee for the high-quality production of cut chrysanthemums. Hyperspectral remote sensing technology has been widely used in precision agriculture due to its rapid, convenient, and nondestructive advantages, but relatively little is known about its use for predicting the PWC of cut chrysanthemums. Therefore, this study aimed to evaluate the performance of hyperspectral reflectance from different leaf layers for estimating the PWC of cut chrysanthemums. A hyperspectral spectroradiometer was used to collect hyperspectral reflectance data (350-2500 nm) from three leaf layers at different critical growth periods. Immediately following the spectra measurements, cut chrysanthemum canopies were sampled for PWC. Spectral index and partial least square regression (PLSR) were then used to establish PWC estimation models of cut chrysanthemums. The results showed that the first leaf layer (LL1) was the optimal leaf layer for estimating the PWC of cut chrysanthemum. The new proposed two-band spectral index, NDVI-LL1 (R2280, R1885), exhibited moderate prediction capability for PWC cut chrysanthemum (R2=0.54658, RMSE=0.02352). Moreover, compared with the spectral index model, the model using the PLSR-LL1 showed the best performance for estimating the cut chrysanthemum PWC (R2=0.93510, RMSE=0.00887). Our results can provide technical support for spectral monitoring of PWC and precise irrigation in cut chrysanthemums.