Research On Grayscale Image Recovery Method In Underwater Scattering Medium Based On Polarization Parameter

In the process of light transmission in the underwater scattering environment,due to the scattering and absorption of particles with wavelengths in the medium,the originally ordered wavefront phase will be severely distorted and the amplitude of the light wave will be attenuated,which results in a dramatic degradation of image quality,significantly reduced image signal-to-noise ratio,and difficulty in accurately capturing and identifying objects.Because polarization imaging system can simultaneously obtain intensity information and polarization information in two-dimensional space,and has the advantages of large field of view,small size,strong real-time performance,light weight,and low system cost,it is widely used in underwater scattering medium imaging.field.In this paper,some key technologies in the grayscale image restoration method of underwater scattering medium are studied based on polarization technology,aiming to further improve the polarization imaging effect in scattering medium under complex environment by using polarization parameters.The main scientific research work of this paper is as follows:1.According to the characteristic that different polarization imaging methods have different effects on different components of light,this paper summarizes the technical solutions of polarization imaging,and summarizes them into differential imaging solutions,backscatter suppression solutions,and non-attenuated target light inversion solutions.Aiming at the problem that the traditional differential imaging method only processes the imaging results from the output dimension,a differential imaging model based on 3×3 Mueller matrix combined with polarization active illumination modulation is established to realize the automatic adaptation of the polarization state of the active illumination and the differential method.A differential imaging method that realizes complete modulation from two dimensions of input and output is proposed.The proposed method has good results in detail preservation and image restoration for target recognition and detection in strong scattering media.Compared with the traditional differential method,the image quality index is increased by 3 times,and the output image quality is significantly improved.2.Aiming at the problems of poor robustness and distortion of restoration results caused by digital computation in some imaging methods,an optimal model of polarization degree of backscattered light was established,and a polarization imaging method based on complete Mueller matrix was proposed to suppress backscattered light directly.The proposed method can realize the suppression of the polarized portion of backscattered light component in the backscattered light when the polarization degree of the backscattered light reaches the optimum.At the same time,the proposed method can be implemented both computationally and by spatial optical path modulation,that is,"what you see is what you get"(the image received by the imaging plane is the final scattering suppression result).In addition,since the method only achieves scattering suppression through polarization optical modulation,the robustness of the method is strong,and the imaging effect can be further improved by combining digital image processing methods.3.Aiming at the problem of incomplete suppression of backscattered light in real scenes,a digital computational polarization imaging method based on virtual illumination is innovatively proposed.The method breaks through the realistic limit by means of computational imaging to achieve the degree of polarization of the backscattered light to be 1,and combines the Mueller matrix information at the background to determine the virtual Stokes vector of the optimal active illumination light source,which theoretically achieves complete suppression of backscattered light.4.Aiming at the problem that the image recovery method based on atmospheric imaging model needs to have a background area in the image to estimate the backscattered light,and also needs to estimate the intermediate variables and cannot get rid of the human-computer interaction,an automatic restoration method of grayscale images in underwater scattering medium without background or any prior knowledge is proposed.The proposed method establishes a physical model based on the idea of intermediate variable degradation,analyzes the physically feasible region of backscattered light,breaks the "point-to-point" correspondence of orthogonal images,and automatically obtains the optimal parameters of backscattered light.Experimental results show that this method has significant advantages in scattering suppression,detail restoration and noise control.In the case of low turbidity,the restoration effect can be close to the naked eye visual effect in the case of air or weak scattering,and the gray range of the image is 1.7 times higher than that of the CLAHE method.5.Aiming at the problem that most image processing methods need to manually adjust the input parameters,a fast polarization imaging method without adjusting any parameters is proposed.The proposed method is practical and simple.It only takes 0.007 seconds to process a 320,000-pixel image,and it has remarkable effects on different scattering scenes such as underwater imaging and image dehazing.