Image Fusion And Enhancement For Dual Camera Mobile Phone

Using mobile phone to take pictures has become the most important source of people's daily acquisition of images,so mobile phone manufacturers began to grab users through mobile phone camera,which directly led to the better and better function of mobile phone shooting.One of the emerging mainstream technologies is using multiple cameras to image.One of the combination schemes of color and mono camera is one of them.This paper studies the fusion and enhancement of color and mono images.The first is the fusion of two images with the same resolution under dark light conditions,and the second is the enhancement of high resolution mono and low resolution color images under normal illumination conditions.In this paper,we mainly improve HIS transform for image fusion under dark light conditions.According to the shortcomings of HIS transform fusion which only aims at Ichannel image enhancement,guided filtering and sparse representation are introduced to enhance all channels of color image,and according to the spectral distortion caused by spatial inconsistency of guided filtering image,a method of using SFIM to transform mono image by each channel of color image is proposed,which can not only maintain the spatial consistency of image,but also preserve the spatial consistency of image.The spectral characteristics can also improve the clarity of the image.Firstly,a method based on registration is proposed to obtain the optimal solution for high and low resolution image fusion under normal illumination conditions.Then,an adaptive weight algorithm is proposed to improve the spectral distortion.Finally,according to the error size,a combination of HIS transform and registration method is used to obtain high detail performance and spectral preservation characteristics.Secondly,deep learning is applied to the fusion of super-resolution,and the idea of deep learning for color image preprocessing and image fusion is proposed,as well as the fusion-based super-resolution model using only depth learning.Finally,two new ideas of super-resolution for single image are proposed.First,based on the imaging characteristics of color images,the color images are arranged based on Bayer filter,and then processed continuously.Finally,the high resolution results are obtained by fusion.The second is to introduce the reference image as input to guide the process of image super-resolution construction.