Spatial Prompt and Wavelet Mamba-Based Multi-Scale Cross-Domain Feature Fusion Network for Segmentation of Mining-Disturbed Land

Segmentation of mining-disturbed land is of great significance for eco-geological environment monitoring. Although existing methods possess strong segmentation capabilities, mining-land exhibits irregular edges, different spatial size, and global texture variability, which lead to difficulties in extracting discriminative features, thereby limiting the accuracy performance. This study first built an RGB-based binary semantic segmentation dataset, covering typical mining-lands in Fujian Province of China. Then a spatial prompt and wavelet Mamba-based multi-scale cross-domain feature fusion network (SWDF-Net) was proposed. (1) Wavelet Mamba-based dual-frequency collaborative enhancement module: high-low frequency information was decoupled by wavelet transform, and further collaboratively enhanced by fusion of multiple local details and Mamba-guided global information. It can highlight mining-lands' high-frequency edge features and low-frequency texture patterns. (2) Cross-domain feature alignment and fusion module: cross-domain statistical calibration and channel conditional modulation were used to narrow spatial-frequency feature distribution gap, which facilitates cross-domain feature alignment and effective fusion. (3) Spatial prompt-based multi-scale feature weighted fusion module: pixel-level weight maps were generated by spatial prior prompt that derived from a spatial decoder and edge gate-based branch, which adaptively fuses former multi-scale dual-domain features. The SWDF-Net achieved best Intersection over Union of 72.22% for mining-land and performed competitively on the ISPRS Vaihingen dataset.