A study of rice phenological development stages estimation from field server images

The accurate rice phenological development stages estimation is important for rice yield estimation, since each phenological developmental stage possesses different increase rates of biomass and leaf area. This paper presents a method to estimate the rice phenological development stages: seedling, tillering, heading, and harvest. This method uses a vegetation index, i.e. Excessive Green (ExG) integration to classify the phenological development stages. The ExG can be calculated from the intensity of different wavelengths in the field server images. The estimated phenological development stages is observed from ExG. Our method is compared to those from KKU rice model simulation and validated with the data obtained in 2013 during the wet season in Roi-Et and during the dry season in irrigated area in Suphanburi. The results show that the estimated phenological development stages obtained from the vegetation index ExG agree well with simulated values from KKU rice model simulation and field observation.