Snapshot Mosaic Hyperspectral Cameras (SMHCs) face challenges in motion perception due to narrow-band spectral filters, which can cause blurry images.
The paper proposes a hardware-software collaboration that combines SMHCs with neuromorphic event cameras to enhance frame clarity and recover spectral details using a new model called spectral-aware Event-based Double Integral (sEDI).
A Noise Awareness training framework, along with a specialized Event-enhanced Hyperspectral frame Deblurring Network (EvHDNet), improves the robustness and effectiveness of image deblurring, showing better results than current top methods.