Estimation of Normal Rice Yield Considering Heading Stage Based on Observation Data and Satellite Imagery

Damage assessment methods for paddy fields are essential to reduce the damage caused by decreases in rice yield. In order to apply agricultural insurance, it is very important to determine the normal yield, i.e. expected yield under normal weather conditions for each paddy field. The objective of this study is first to develop a model to estimate the yield of paddy based on satellite images, and finally to estimate the normal yield of each plot. In this study, multiple spectral information such as visible and near-infrared bands extracted from Sentinel-2 images and several indices were used to perform multiple regression analysis considering the heading stage based on NDVI. The results of this equation (RMSE = 1.22 t/ha) were used to estimate the normal yield for 2020. The results suggest that satellite data can be used effectively to estimate yield, and used to estimate the normal yield of each paddy plot. In the future, it is expected to contribute to the quantification of losses for the calculation of compensation amounts for agricultural insurance.