Multichannel Imaging Restoration Based On Simple Lenses

Because of optical aberration,the multi-channel imaging with a simple lens will show serious blur.In order to solve this problem,the modern lens optical imaging systems usually contains dozens of lenses,using the positive and negative offset aberrations of different lenses to counteract each other,and ultimately enhance the image quality.But this method causes imaging system to be complex design,huge,and high cost.Many applications,such as remote sensing photography,combat reconnaissance,national defense monitoring and so on,all have an urgent need for portable simple lens imaging systems,because this small,lightweight imaging system can be loaded and used more easily.Therefore,an image restoration method based on a simple lens is proposed in this paper,which is to obtain high quality images by the image restoration process of the blurred images rather than complex lenses.The concrete work of this paper is:(1)For the RGB fuzzy image taken by a simple lens,propose a restoration algorithm combined with adaptive sparse gradient and inter channel correlation.The algorithm uses the sparse constraint smooth area and edge points for different fuzzy images of the three color channels,and uses the correlation of the three channels to help restore the edge information of one channel through the others' edge information.The simulation experimental results show that the proposed method has better restoration quality and better visual effect for the blurred image obtained by the simple lens.(2)For the multispectral image taken by a simple lens,propose a restoration algorithm based on the guide image's gradient.The algorithm uses inter channel correlation of multispectral images(multispectral images can be approximated by linear combination of less base images)to calculate every guiding image of blurred images,then uses the gradient similarity constraint of the guiding image and the original image and gradients sparseness constraint to estimate the sharp image.The simulation experimental results show that the proposed algorithm improves the restoration quality and visual effect of defocus blurred multispectral images.