An Image Enhancement Algorithm For Turbid Underwater Image Based On Multiple Methods

Underwater resource development is more and more important, but some underwater operations, such as scientific exploration and construction work depending on underwater observation, are with visual difficulties because of the turbid water. To improve underwater visual quality, we propose a comprehensive image enhancement algorithm in this paper, which consists of four sections. The first section is Bilateral Filter used to reduce image noise, then the image is dehazed based on Dark Channel Prior (DCP). Thirdly, we propose one White Balance method to adjust colors in underwater images. At the end of the algorithm, Contrast Limited Adaptive Histogram Equalization (CLAHE) is adopted to improve the contrast in images. Besides, to verify this algorithm, a series of underwater image experiments are conducted on the turbid water test platform. More than 300 turbid underwater images are captured and some of them are processed through the proposed algorithm. By using Image Quality Evaluate Metrics and comparing degraded images with result images, we evaluate the algorithm in subjective and objective aspects, and conclude that the proposed algorithm can comprehensively improve the quality of images with better color, contrast and more details, which has a great significance to underwater operations.