Research On Image Enhancement Based On Structure Preserving

With the rapid development of computer vision technology and multimedia technology,people have higher requirements for the visual effects of color images.Image enhancement technology based on structure preservation is an important part of image processing.Its purpose is to preserve and highlight important features of images such as texture details,layering,and contrast,so as to obtain high-quality result images.Among them,color transfer and underwater image enhancement are research hotspots in recent years and have received widespread attention.At this stage,the traditional color transfer method has problems such as loss of texture details,insufficient sense of hierarchy,inaccurate semantic region segmentation,and color misrepresentation,while the traditional underwater image enhancement method has problems such as loss of texture details,obvious color deviation,insufficient contrast,and supersaturation.In response to the above problems,the color transfer method and underwater image enhancement method are studied in depth with the clue of structure preservation.The main research contents and innovations are as follows :(1)Aiming at the problems of texture detail loss and lack of hierarchy in traditional color transfer methods,this paper proposes a new color transfer method based on intrinsic decomposition and frequency tuning theory(IFCT method).Firstly,the reference image and the source image are divided into a reflection image containing color information and an illumination image containing texture information by intrinsic decomposition.Then,color mapping is performed only between the reflection images to reduce the loss of texture details of the resulting image caused by color transfer.Secondly,a new saliency detection strategy is designed based on the frequency tuning theory.The reflection image of the reference image and the source image is divided into saliency region and non-saliency region,and then combined with the multidimensional Radon transform to perform color transfer in the corresponding region,so that the result image maintains the layering of the source image.Finally,the color transfer result of the reflection image is combined with the illumination image of the source image to obtain the result image.A large number of experimental results show that,The proposed IFCT method can not only effectively maintain the texture details and layering of the resulting image,but also have better visual effects and better performance in subjective evaluation and objective evaluation.(2)Aiming at the problems of texture detail loss,obvious color deviation,insufficient contrast and oversaturation in traditional underwater image enhancement methods,a new underwater image enhancement method is proposed based on the white balance theory of multi-scale edge preservation and red-blue channel compensation(MWIE method).Firstly,the improved white balance strategy is used to correct the color of the underwater image,and the contrast preserving algorithm is used to improve the visual effect of the result image.Secondly,a multi-scale edge preserving algorithm is designed based on the weighted least squares theory to make the result image better preserve the edge texture details.Finally,the homomorphic filtering algorithm is used to filter the image to avoid the problem of oversaturation,and finally the result image is obtained.A large number of experimental results show that,the proposed MWIE method can not only effectively maintain the texture details of the result image,but also solve the problems of color deviation,insufficient contrast and overall quality of the result image obtained by the proposed method is better.