Design of Power Efficient Saturation Based Image Dehazing Algorithm for Haze Removal
Pa Saisivakumaran, S Surendhar, V Saai Miruthunjayan, Karuthapandian Mariammal · 2024
Foggy weather can significantly impact the sharpness of images captured by real-time systems like object identification, remote sensing, and surveillance. To address this, haze-removal systems are utilized, with a modified version of a saturation-based image dehazing algorithm, along with techniques like histogram equalization and guided image filtering, being implemented. The proposed dehazing system's VLSI architecture consists of seven pipelined stages. A 15x15 window minimum filter is used to down-sample the foggy images to estimate atmospheric light. Additionally, a saturation-based transmission map, which employs a patch-based approach rather than a pixel-based one, is used for estimation. Unlike other patch-based methods, this approach excludes edge detection and image filtering units to avoid halo artifacts around edges. The hardware implementation of these algorithms has been performed, yielding the corresponding outputs.