Monitoring Cropland Abandonment in Hilly Areas with Sentinel-1 and Sentinel-2 Timeseries

Abandoned cropland may lead to a series of issues regarding the environment, ecology, and food security. In hilly areas, is prone be abandoned due scattered planting, relatively fewer sunlight hours, lower agricultural input–output ratio. Furthermore, impact rainfed differs from irrigated cropland; thus, corresponding land strategies vary accordingly. Unfortunately, monitoring still an enormous challenge in areas. this study, new approach was proposed by (1) improving availability Sentinel-1 Sentinel-2 images processes, (2) obtaining training samples multisource data overlay analysis timeseries viewer tool, (3) mapping annual cover all available images, samples, random forest classifier, (4) spatiotemporal distribution areas assessing land-cover trajectories along with time. The result showed that had F1 scores (0.759 0.8) compared (0.836 0.879). High overall accuracies around 0.90 were achieved, kappa values ranging 0.851 0.862, which outperformed existing products accuracy spatial detail. Our study provides reference for extracting