Dynamic Frequency Feature Fusion Network for Multi-Source Remote Sensing Data Classification
This work addresses the problem of land cover classification for remote sensing applications, presenting an incremental improvement over existing methods.
The paper tackles the challenge of multi-source remote sensing data classification by proposing a Dynamic Frequency Feature Fusion Network (DFFNet) that dynamically learns frequency domain features and fuses cross-modal data, achieving state-of-the-art performance on benchmark datasets.
Multi-source data classification is a critical yet challenging task for remote sensing image interpretation. Existing methods lack adaptability to diverse land cover types when modeling frequency domain features. To this end, we propose a Dynamic Frequency Feature Fusion Network (DFFNet) for hyperspectral image (HSI) and Synthetic Aperture Radar (SAR) / Light Detection and Ranging (LiDAR) data joint classification. Specifically, we design a dynamic filter block to dynamically learn the filter kernels in the frequency domain by aggregating the input features. The frequency contextual knowledge is injected into frequency filter kernels. Additionally, we propose spectral-spatial adaptive fusion block for cross-modal feature fusion. It enhances the spectral and spatial attention weight interactions via channel shuffle operation, thereby providing comprehensive cross-modal feature fusion. Experiments on two benchmark datasets show that our DFFNet outperforms state-of-the-art methods in multi-source data classification. The codes will be made publicly available at https://github.com/oucailab/DFFNet.