REOUN: restoration and enhancement of optical imaging underwater based on non-local prior

The complex aquatic environment of underwater imaging sometimes leads to significant image distortion. Due to light absorption and dispersion in aqueous medium, underwater photographs frequently experience serious quality loss, including poor visibility, contrast reduction, and colour divergence. Reducing colour cast, increasing contrast, and improving visibility in these photographs is a difficult task. To increase the quality of underwater photos, restoration and enhancement based on Non-Local previous technique has been developed. However, the utilisation of several undersea image restoration and enhancement techniques is hampered by the over- or under-enhancement they yield. Initially, the use of Non local prior averages pixel intensities of various locations spread across the hazy image plane have a quasi-linear association with those over the equivalent haze free images. Secondly, a dual optimization function is employed to reduce the size of the solution space to remove blur. Furthermore, by estimating the imaging parameters using the information present in the full image, this unique dual optimisation technique produces a more dependable restoration. Thirdly, the negative interference brought about by the region segmentation is removed by the use of a gradient filter. Finally, an improved weighted grey edge method is adopted to enhance image brightness and visibility. A comparison is made between the REOUN and an existing methods in terms of both objective and subjective visual impact. We compared the edge information of the restored results since texture and details are crucial to images and serve as evaluation criteria to gauge an image's performance. Comparison results show that the suggested REOUN retains the edges best over other techniques.