Cropland Data Extraction in Mekong Delta Based on Time Series Sentinel-1 Dual-Polarized Data

In recent years, synthetic aperture radar (SAR) has been a widely used data source in the remote sensing field due to its ability work all day and weather conditions. Among SAR satellites, Sentinel-1 is frequently monitor large-scale ground objects. The Mekong Delta major agricultural region Southeast Asia, so monitoring cropland of great importance. However, it challenge distinguish from other objects, such as aquaculture wetland, this region. To address problem, study proposes statistical feature combination dual-polarimetric (dual-pol) time series based on m/χ decomposition method. Then put into proposed Omni-dimensional Dynamic Convolution Residual Segmentation Model (ODCRS Model) high fitting speed classification accuracy realize extraction Experiments show that ODCRS model achieves an overall 93.85%, MIoU 88.04%, MPA 93.70%. results our method can effectively areas wetlands.