Research Of Underwater Image Restoration Method

The digital image restoration technique is applied to many fields,such as microscopic imaging,remote sensing images,medical images and so on.Because of the complexity of the underwater environment,the study of underwater image restoration is lack.With the development of society and science technology,water pollution is more and more serious,it is one of the important issues of concern that how to solve the problem of water quality.More and more application requires to be able to get high quality images such as water conservancy project,environmental monitoring,underwater topography under recovery.Considering of the complexity of the underwater environment and the cost of the hardware device,the underwater image restoration of software technology becomes more important.The main work of this paper is as follows:1,analysis of blur types and characteristics of underwater imageThe complexity of the underwater environment will lead to the differences in the types of blurred image,blur types classification can be divided into three class:motion blur,Gaussian blur and defocus blur.In this paper,method of underwater image recognition based on spectrum analysis,we study spectrum analysis of underwater image blur identification method by directly Fourier transform.According to the spectrum that is the Fourier transform image respectively after normalized logarithmic,frequency shift,image smoothing,image enhancement.We analysis spectral characteristics in different types of blur image.So that different degradation models are used to make the corresponding estimated recovery.2,the establishment of underwater image restoration based on turbulence-degraded modelBecause of the water and air are all the fluid medium,underwater image degradation is equivalent to the degraded images under severe atmospheric disturbance.It is proved that the underwater image degradation model is proved to be feasible.Therefore,according to the degradation of the turbulence model,and theblur function type is determined,the underwater image restoration based on turbulence can be equivalent to the parameter estimation.3.the estimation of parameter k for turbulence-degraded modelBased on the turbulence-degraded model,the classical method is to estimate the parameters k of the blur function by the image quality assessment.This paper designs a method to solve the optimal value according to the blur-MSE cost function.A search strategy from coarse to fine for underwater degraded images is presented to estimate parameters k.According to the estimated optimal value,the blur function can be reconstructed.Because the research of this paper focuses on the estimation of the blur parameters,we use the simple Wiener filter for non-blind restoration.4,the establishment of blind underwater image estimation based on alternating iterationDue to the underwater complex environment,in the process of underwater imaging,underwater image degradation is only regarded as the turbulence model is not able to meet the needs of the actual.In this paper,the underwater imaging system is considered as a linear image degradation system,and the blur function and the underwater image are estimated by alternating iteration.Although this method can't achieve the removal of turbid particles in the water,and the influence of the underwater light,the alternating iteration method can effectively improve the visual effect and detail information of underwater images.