Research On Image Restoration In Inhomogeneous Medium

Because the medium(fog,haze,water etc.)causes interference in transmission process,image contrast is low,colors become dark,and texture is fuzzy and other phenomenons.The degradation of fog images and underwater images is mainly due to the backscattering of suspended particles,then the image is blurring,the original color of the object surface is covered,and the local area of the object is not clear.They affect the performance of the algorithm in the field of computer vision,such as object tracking,scene analysis,scene reconstruction and so on.The reliability has been seriously reduced.Therefore,the image restoration technique in inhomogeneous media propagation has an important application value.The clearness methods in fog and underwater image of inhomogeneous media propagation are mainly divided into image enhancement method and image reduction method.The degraded image enhancement method does not depend on the specific model,and emphasizes the information of the interest region and suppresses the feature of non interest;the degraded image restoration method relies on the specific model,and fully analyzes the root cause of image degradation,and restores it to the state before image degradation.The main work and research contents include the following aspects:First,this paper introduces attenuation and scattering process of light in inhomogeneous medium,briefly expounds the effects of attenuation and scattering for the imaging process.Then image fusion technology and some the classical degraded image enhancement algorithms are described briefly.Secondly,the restored image presents the halo effect and dim phenomenon,so optimized dehazing method based on guided filter and image fusion is proposed.Based on the characteristics of the region where the atmospheric light is located,the brightness threshold range and the flat threshold value of the atmospheric light are set,and the atmospheric light position is determined.The edge detection algorithm is used to obtain initial estimation of medium transmission.The coarse estimation transmittance is estimated by Gaussian filter.The image fusion between the coarse estimation and initial estimation of transmission is performed.Then,the dehazing image is get.Finally,underwater image enhancement method is easy to make the image distortion,color oversaturation.And image color becomes dim after traditional image restoration method is used to underwater images,so an underwater image restoration method based on dark channel and image fusion is proposed.The establishment of underwater optical imaging model is based on the characteristics of underwater imaging.Because underwater optical imaging model is similar to the atmosphere model,the global light is corrected.The underwater restoration image is obtained by the optimized dark channel prior algorithm.The histogram equalization method is applied to underwater image enhancement.The image fusion between the underwater restoration image and enhancement image is performed.