A universal estimation model of fractional vegetation cover for different crops based on time series digital photographs

The green fractional vegetation cover (FVC) is an important parameter in monitoring crop growth and predicting aboveground biomass. In this study, we monitored crop growth with digital cameras installed at four automatic weather observation stations in different parts of China, from 2010 to 2016. With each station having a particular type of crop, nine color vegetation indices were calculated from the acquired time series digital photographs to arrive at an FVC estimation model applicable to sugarcane, maize, cotton and paddy rice. For individual crop types, our results show that the Excess Green (ExG) is the optimal color vegetation index for the estimation of sugarcane FVC, the Normalized Difference Index (NDI) is the optimal color vegetation index for the estimation of maize FVC, and the Vegetative (VEG) color vegetation index is optimal for the FVC estimation of cotton and paddy rice. However, owing to its higher coefficient of determination (R2), and lower root mean square error (RMSE) and mean absolute error (MAE) of 0.9504, 0.0721 and 0.0545, respectively, the Color Index of Vegetation Extraction (CIVE) is found more universally applicable for FVC estimation of the four crop types under investigation. The CIVE index has therefore been proposed in this study to be optimal for FVC estimation in sugarcane, maize, cotton and paddy rice mixed cropping agro-systems which are especially common in small and highly fragmented agricultural landscapes such as those in urban and peri-urban areas.