Feature Fusion and Difference Enhancement Optimization Network for Remote Sensing Image Semantic Change Detection

Semantic change detection (SCD) presents significant challenges in the realm of remote sensing image (RSI) analysis, focusing on the detection, localization, and examination of pixel-level land cover changes captured over time via multitemporal remote sensing images. Although various methods have been suggested in previous studies for remote sensing change detection, the issue of effectively identifying changes in finegrained details remains an area needing further investigation. To tackle this problem, we introduce the Feature Fusion and Difference Enhancement Optimization Network (FFDEONet), which combines feature fusion and difference enhancement optimization modules specifically designed for SCD in remote sensing images. Results from experiments conducted on the SECOND dataset validate the effectiveness and advantages of the FFDEONet approach in performing SCD tasks within remote sensing imagery.