An attention mechanism based infrared imaging detail enhancement algorithm design and implementation with FPGA

This article proposes an adaptive infrared image detail enhancement algorithm guided by visual attention map for the issue that the output image of infrared detection equipment features concentrated gray value and low contrast. The algorithm separates the base layer and the detail layer via edge-preserving filters such as guided filtering, compresses the dynamic range of the base layer image, and adaptively enhances the detail layer image. Subsequently, the low contrast infrared image is reconstructed by the base layer after dynamic range compression and the detail layer after detail enhancement. Finally, the algorithm is deployed on the FPGA platform, which significantly improves the image contrast and meets the real-time requirements of the mid-wave infrared 640×512 array detection system.