Research On Underwater Image Restoration Algorithm Based On Suppression Of Non-uniform Scattering

Marine pasture is fast developing in China.Advanced underwater on-site observation and measurement equipment is an important guarantee for realizing highefficiency bottom sowing multiplication and precise management.It is of great significance to use underwater cameras to replace human beings to enter dangerous areas,and to replace divers in various complex sea areas to conduct on-site real-time observation of various resource organisms.However,due to the influence of water absorption,scattering,sediment,lighting,etc.,the pictures directly taken by the camera in the underwater environment often have low contrast,lack of detail,and serious loss of high-frequency information.The overall image quality is low.In the underwater environment,scattering is the main factor causing the degradation of image quality.Compared with the uniform scattering in the air medium,the water in the ocean pasture is complex and changeable,and non-uniform scattering will occur in the underwater environment of different sea areas.The existing underwater scattering suppression algorithms are difficult to apply.Therefore,in order to achieve underwater image restoration in a non-uniform scattering environment,this thesis has researched on the suppression of non-uniform scattering effect on underwater image.The main research content and innovations of the paper can be summarized as follows:(1)This article first discusses the propagation characteristics of visible light in water based on the light scattering model.In addition to separately studying the absorption,forward scattering,and back scattering that light will encounter when traveling underwater,two types of non-uniform scattering are elaborated in detail.Firstly,the phase equation of Mie scattering is discussed.The phenomenon of nonuniform distribution of scattered radiation in turbid water environment.Secondly,the phenomenon of color scattering is discussed.The wavelength selectivity of water molecules and colored underwater suspended solids will cause the non-uniform radiation intensity of light of different wavelengths scattered by the water medium,and color scattering of sea water.The existing underwater scattering suppression algorithms have unsatisfactory effects on the improvement of the image quality reduction caused by the above-mentioned two kinds of non-uniform scattering.(2)For the color scattering caused by the inconsistent radiation intensity of light of different wavelengths scattered by the solid in water,we propose a two-step method to suppress it.First,the overall color shift of the image caused by color scattering can be corrected by compensating for the missing color information of the color channel.Then,the modified method for calculating transmission rate based on L0 algorithm is applied to enhance contrast of underwater image.(3)In turbid water,as the concentration of suspended particles in the water increases and the size increases,the proportion of Mie scattering becomes larger and larger,and the scattered radiation tends to concentrate in the direction of 90 degrees,and the distribution of scattering becomes more and more non-uniform.Moreover,turbid water can also cause problems such as extremely low image signal-to-noise ratio,severe forward scattering and blur image details.Traditional scattering suppression methods are mostly based on the Mc Cartney scattering model derived from Rayleigh scattering,so the turbid underwater environment dominated by Mie scattering poses a severe challenge to these traditional image processing algorithms.In this article,by analyzing the reasons for the uneven distribution of turbid water scattering radiation,we proposed a neural network-based method to replace expensive and complex special imaging equipment to decrease the influence of uneven scattering on image.Experiments prove that the method in this paper reduces the noise,enhances the edge definition and overall contrast of the image,and achieves a better scattering noise suppression effect than the traditional method in a turbid water environment.(4)This article clarifies that the further work is to develop methods suitable for suppressing the scattered noise in highly turbid waters with turbidity(NTU)greater than 50.The absorption of dense suspended solids in high turbidity water is extremely serious,and additional lighting are needed to enhance the reflected signal of the target object to overcome the absorption.However,the intensity of scattered light is positively correlated with incident light.As the illumination light increases,the scattering caused by suspended particles will become more intense,making the image more blurred and the imaging effect worse.Therefore,for future applications in high turbidity,this paper proposes an approach based on the single-channel scattering suppression algorithm and the near-infrared lighting method.Experiments prove that this method has the potential to obtain clear images in highly turbid water.