Estimation of Anthocyanins in Leaves of Trees with Apple Mosaic Disease Based on Hyperspectral Data

Anthocyanins are severity indicators for apple mosaic disease and can be used to monitor tree health. However, most of the current studies have focused on healthy leaves, few estimated anthocyanin content in diseased leaves. In this study, we obtained hyperspectral data leaves with disease, analyzed spectral characteristics different degrees Mosaic constructed screened index sensitive content, improved estimation model. To improve conciseness model, integrated Variable Importance Projection (VIP), Partial Least Squares Regression (PLSR), Akaike Information Criterion (AIC) select optimal PLSR model its independent variables. Sparrow Search Algorithm-Random Forest (SSA-RF) was accuracy. Results showed following: (1) increased gradually aggravation disease. The reflectance blade spectrum visible band increased, red edge moved short wave, phenomenon “blue shift spectrum” occurred. (2) VIP-PLSR-AIC selected 17 variables from 21 indices. (3) Variables were construct PLSR, Back Propagation (BP), Support Vector Machine (SVM), Random (RF), SSA-RF estimate content. accuracy stability better than other models. set determination coefficient (R2) up 0.955, which is 0.047 higher that RF 0.138 SVM lowest at leaf scale provide a reference studies, including theoretical basis large-area, high-efficiency, high-precision monitoring mosaics using remote sensing technology.