Detecting Paddy Field Flooding Using Sentinel-1 Satellite Data and Paddy Parcel Boundary Data

Abstract

The required amount of irrigation water in paddy fields varies depending on the time of year, because water intake permitted from rivers for water rights is determined by period. Generally, water intake permitted for a period from soil puddling is the largest. Periods for soil puddling have been advanced on a trial basis in response to changes in farming programs. When the periods for soil puddling are advanced, surveys on how early soil puddling was actually done in paddy fields are performed. The purpose of this study was to clarify whether synthetic aperture radar (SAR) data from the Sentinel-1 satellite can be effectively used in such surveys. Accuracy in determining whether each paddy field was flooded on the satellite observation date using Sentinel-1 satellite data and paddy parcel boundary data was verified. The average pixel value (backscattering coefficient) of the SAR image was equal to or less than the threshold value used to determine whether a paddy field was flooded. Verification of accuracy was performed in survey areas where the ground surface was observed by both the Sentinel-1 satellite and the Sentinel-2 satellite on the same clear day (20 April 2018). In addition, highly accurate results were obtained by determining whether each paddy field was flooded using multispectral data of the Sentinel-2 satellite and paddy parcel boundary data as verification. Determination accuracy was 79.1% when the VV-polarized SAR image was used and 75.8% when the VH-polarized SAR image was used. Therefore, VV-polarized SAR images are more suitable than VH-polarized SAR images for determining whether paddy fields are flooded. In addition, Sentinel-1 satellite data can be used effectively in surveys.