Detection of Paddy Fields in Maharashtra’s Eastern Region During the Kharif Season Using Sentinel-1A SAR Data and Geographic Information System (GIS)

For the effective harvest of rice, Rice (Oriza sativa) fields must be detected and monitored. Remote sensing information from optical and radar systems should be used for the monitoring, it has the ability to cover a vast geographical area. Cloudy days are common in tropical regions in monsoon (Kharif) season therefore, using optical satellite data to locate paddy fields might be difficult. Application of radar remote sensing data to detect paddy fields in cloudy condition can solve the problem. The study was carried out to explore the capabilities of sentinel-IA temporal backscatter at vertical transmitted & vertical received (VV) and vertical transmitted & horizontal received (VH) polarizations to detect paddy fields based on rice crop phenology. The data duration is selected based on paddy growing season i.e., Kharif (June-October, 2021) in the Bhandara district, eastern region of Maharashtra. The satellite data downloaded and pre-processed in Google Earth Engine (GEE) platform. Temporal sentinel-IA composite image and taking training samples of different phenological phases of rice plants to detect paddy fields. The SNAP tool’s random forest (RF) method is used for supervised classification, while the ground observation points serve as validation. According to the study, sentinel-IA temporal backscatter at vertical transmitted & vertical received (VV) and vertical transmitted & horizontal received (VH) polarization allows for the detection of paddy fields in Kharif season with an overall accuracy obtained from 77% and 86% respectively.