π°οΈ New AI method merges radar and optical satellite images
Researchers introduced a deep neural network framework to fuse heterogeneous radar and optical remote sensing imagery. This work was conducted by teams from various institutes, published on September 19, 2026. The method achieved 91.38% overall accuracy and a 0.892 Kappa coefficient on the SEN1-2 benchmark. It successfully surpasses ten competitive baselines while keeping inference cost low at 11.4 ms per sample. Future work will likely focus on applying this robust fusion technique in diverse remote sensing applications. π°οΈ