Underwater Image Enhancement Algorithm Based On Super Resolution

With the increasing population of our country,there is a constant shortage of resources on land,and people are paying more and more attention to marine resources.Underwater vision technology plays an important role in the development of marine resources,environmental monitoring,exploration,and scientific investigation.Underwater fishing robot acquires images in underwater environment.Because the environment is too complicated,natural light is incident on the water,and the underwater image caused by the absorption,refraction and scattering of light causes color shift,detail blur and low contrast.The human visual system can not meet the machine target recognition and fishing needs of underwater fishing robots,which brings great difficulties for the accurate identification and stable capture of our subsequent underwater fishing robots.In order to improve the accurate recognition and stable capture of marine products by underwater fishing robots,in order to solve the problems of color shift,detail blur and low contrast in underwater images,a super-resolution underwater image is proposed.The enhancement algorithm takes two steps,namely color correction,detail enhancement,etc.,to make the underwater image clear.The work done in this article includes the following sections:(1)This paper adopts a new framework for enhancing the phenomenon of the degradation of the original underwater image quality.A color correction algorithm including multi-scale fusion and a multi-scale residual-dense super-resolution generation countermeasure network(MSDR-SRGAN)algorithm to make underwater images clear.(2)For the color shift and low contrast phenomenon existing in the original underwater image,the underwater image color correction method based on multi-scale fusion is used.Compared with other commonly used color correction methods,the color correction method can well remove the color shift phenomenon of underwater images,and is more suitable for underwater images.(3)For the underwater image after color correction,there are such phenomena as blur and detail loss.From the perspective of image super-resolution,the super-resolution generation confrontation network is improved,and a multi-scale residual-dense super is proposed.Resolution Generation Against Network(MSDR-SRGAN)algorithm.This method makes the underwater image texture and details clearer,and the final enhanced underwater image is obtained.The experimental comparison shows that the proposed algorithm can effectively solve the phenomenon of color cast,missing detail,blur and low contrast in underwater images,and improve the visual effect of underwater images.After comparing several other commonly used underwater image enhancement methods and the accuracy of target recognition for underwater fishing robots,the effectiveness of the underwater image enhancement method proposed in this paper is verified.