ABSTRACT
The underwater image enhancement is a challenging for the low signal-to-noise ratio and uneven illumination under water. To solve this problem, this paper proposes a novel image enhancement algorithm. First of all, the knowledge representation of the underwater image is constructed by rough set. Then the image is divided into optimal partition according to the approximate equivalence relation between upper and lower approximation. Finally, an improved affine shadow formation model is used to enhance each layer of the image based on the optimal partition. The simulation results verify the effectiveness of the algorithm.
