DFFAN: Dual Function Feature Aggregation Network for Semantic Segmentation of Land Cover

Analyzing land cover using remote sensing images has broad prospects, the precise segmentation of is key to application this technology. Nowadays, Convolution Neural Network (CNN) widely used in many image semantic tasks. However, existing CNN models often exhibit poor generalization ability and low accuracy when dealing with To solve problem, paper proposes Dual Function Feature Aggregation (DFFAN). This method combines context information, gathers spatial extracts fuses features. DFFAN uses residual neural networks as backbone obtain different dimensional feature information through multiple downsamplings. work designs Affinity Matrix Module (AMM) each map Boundary Fusion (BFF) fuse an determine location distribution image’s category. Compared methods, proposed significantly improved accuracy. Its mean intersection over union (MIoU) on LandCover dataset reaches 84.81%.