作者: Hassan Bazzi , Nicolas Baghdadi , Mohammad El Hajj , Mehrez Zribi , Dinh Ho Tong Minh
DOI: 10.3390/RS11070887
关键词:
摘要: This study proposes an effective method to map rice crops using the Sentinel-1 SAR (Synthetic Aperture Radar) time series over Camargue region, Southern France. First, temporal behavior of backscattering coefficient 832 plots containing different crop types was analyzed. Through this analysis, cultivation identified metrics derived from Gaussian profile VV/VH (3 metrics), variance (one metric), and slope linear regression VH metric). Using metrics, were mapped through two approaches: decision tree Random Forest (RF). To validate accuracy each approach, classified compared available national data. Similar high overall obtained both approaches. The a simple reached 96.3%, whereas 96.6% RF classifier. therefore, provides yet precise powerful tool paddy areas.