Remote estimation of carotenoid/chlorophyll ratio in vertical layers using canopy multi-angle spectral data

Precise estimation of the ratio of carotenoid to chlorophyll (Car/Chl) in crop is very important in precision agriculture and early detection of stress. Car/Chl value in leaves is considered as an indicator of stress, especially in the middle-layer and bottom-layer of crop. Traditional remote estimation of Car/Chl was based on the data obtained from the nadir. However, there are few studies focusing on the monitoring of Car/Chl in middle-layer and bottom-layer of crop. The objective of this study was to accurately estimate the Car/Chl in vertical layers, especially in the middle-layer and bottom-layer, using in-situ canopy multi-angle spectral data. We took the winter wheat as the research target, the spectral reflectance at nadir, ±20°, ±40°, ±60° viewing angles and leaf carotenoid and chlorophyll content of the bottom-layer, middle-layer and upper-layer measurement were measured, respectively. Two types of models, vegetation indices (VIs) and neural neural network (NN) were employed. It was concluded that -20° observing angle performed best for Car/Chl estimation in the middle-layer, with highest R² of 0.68 for the MTCI model, and 0.86 for the ANN model, respectively. As for the bottom-layer, the NN achieved better results than the PRI model, corresponding R² values were 0.76 and 0.60, and viewing angles were nadir and +40°, respectively. Additionally, the NN and SIPI models also performed well for Car/Chl assessment in the upper-layer, corresponding viewing angles were nadir and +60°. The results support the use of multi-angle observations for a quick crop Car/Chl estimation in vertical layers and contribute to the precise fertilizer and production management.