The Research And Application Of Underwater Image Enhancement Algorithm

The underwater world contains a wealth of natural resources. Due to the increasing shortage of resources and the growing environmental problem, people gradually turn their attention into underwater world. A comprehensive grasp of underwater world is strategic important and will provide strong guidance to make use of underwater resources more reasonably. Underwater optical imaging is one of main ways to obtain information of underwater world. However, as a result of light scattering and absorption in water, underwater image usually degrades seriously with poor visibility, low contrast and color distortion compared with normal image. The main purpose of this thesis is to enhance the quality of degraded image, improve the visual effects and highlight the useful information of image through underwater image processing. The main work of this paper are as follows:First, the light transmission characteristics underwater are analyzed in detail, making it clear that the image degradation is the consequence of light scattering and absorption. Based on the propagation characteristics of light underwater, the underwater image model is introduced, which offers a theoretical foundation for the subsequent contents of this paper. Then several frequently-used methods for underwater image enhancement are introduced. The advantages and shortages of the algorithms are showed through experiments.Second, the enhancement method based on dark channel prior and color correction is proposed in this paper for the problem of serious degradation of underwater image. The wrong estimation of background light is improved by combining the saturation map. Then the transmission map of red channel is estimated through dark channel prior. Two control factors are introduced to estimate the ratio of attenuation coefficients of different wavelength and the other two transmission maps are obtained through the ratio of attenuation coefficients. The scene radiance is restored based on the estimated background light and transmission maps. Because the restored image is still color casted, the color correction method is applied to correct the shift color. The results show that the proposed algorithm can effectively improve the quality of underwater image and enhance the visual effects.Third, for the easier processing of underwater image and better use of advantages of different algorithms, the underwater image processing software is developed. The software is developed by C++. The interface is designed with MFC and the algorithms in the software are realized with OPENCV to simplify the development process. The software integrates the function of image viewing and editing, image enhancement using histogram transforming, Retinex and Dark Channel Prior.