Image correction and enhancement algorithm for precise positioning of underwater soft manipulator

Vision-based precise positioning underwater presents a major challenge in the field of soft manipulator. A critical hurdle is the harsh underwater environment, which affects the effectiveness of the vision sensor. Our study proposes a novel approach based on refractive geometry for the correction of image distortion in underwater environments. Furthermore, the image enhancement process incorporates three distinct algorithms: the DCP (Dark Channel Prior) algorithm for image enhancement, the ACE (Automatic Color Equalization) algorithm for color correction, and Wiener filtering to effectively suppress noise in underwater images. The efficacy of the approach was tested in the context of a circle-reference tracking underwater task. The experimental results obtained provide a robust validation of the efficacy of the proposed localization approach.