Underwater Image Sharpening Enhancement Research

Underwater captured images often suffer from poor visibility caused by multiple degradations.This not only seriously reduce the visual quality of the image,but also makes it difficult to extract information,which affects the follow-up project.Therefore,using the image processing techniques to enhance degraded image is of great practical significance.To improve the image quality and highlight the vital scene features,the main works of this paper are as follows:(1)First,we propose a hybrid framework for underwater image enhancement,which unifies underwater white balance and variational contrast and saturation enhancement.the improved underwater white balance(UWB)algorithm is integrated with histogram stretching and adaptive color compensation,aiming to better uniform the attenuation difference along the propagation path and remove undesired color castings.On the other hand,a variational contrast and saturation enhancement(VCSE)model is developed based on the enhanced result obtained from UWB.The advantages of VCSE model lie in the improvements of contrast and saturation as well as the elimination of hazy appearance induced by scattering.Moreover,we adopt a fast Gaussian pyramid-based algorithm to speed up the solving of VCSE model.Extensive qualitative and quantitative assessments demonstrate that the proposed approach obtains high quality outcomes,which outperforms several state-of-the-art methods.Application tests further verify the effectiveness and broad application prospects of our proposed method.(2)Aiming at the problems of low contrast,color distortion and noise in underwater image,we additionally proposed an enhancement method based on adaptive color and contrast enhancement and denoising.Firstly,we adopt the Gaussian filter and Bilateral filter to obtain high frequency detail image and low frequency color image respectively.The soft-thresholding operation is subsequently utilized to suppress the noise in the highfrequency component.Accordingly,the low-frequency component is enhanced by constructing an adaptive color and contrast enhancement(ACCE)framework.The ACCE can effectively improve the clarity of image without valuable information loss or overenhancement.Through subjective and objective evaluation,we demonstrate the effectiveness and superiority of our strategy.Finally,the application test verifies the feasibility of the proposed method to enhance other degradation scenes,such as atmospheric hazy images and sandstorm images.