Research And Implementation Of Image Enhancement Algorithm Based On Retinex

Image enhancement is a significant branch of digital image processing field, it's the method which transforms some information or data of images through some kind of strategies to highlight the areas of interest or suppress the secondary information, in order to match the human visual response. In recent years, along with the increasing demand for images in various special application environment such as aerial images, remote sensing images and biomedical images, more and more scholars begin to focus on the image enhancement research. In the process of collecting images, it often dues to the scene conditions, weather, light or image acquisition equipment factors to cause some problems such as uneven illumination, image dynamic range compressed and low contract, so that it can't meet the practical application. While Retinex image enhancement algorithm put forward by Edwin.Land has excellent effect when solving this kind of problem.This paper proposes a new improved algorithm on the basis of in-depth study of several classic Retinex algorithms, in order to solve the problem when applied to low illumination environment that images exist color distortion, slow speed. First according to the algorithm has a risk of changing images' color, the Lab color space is introduced into the Retinex, it separates the luminance and chrominance information, the improved algorithm only processed on the luminance channel to avoid the threechannel's ratio imbalance problem which caused by not distinguishing between them in RGB space, this reserves the color information. Then regarding the time-consuming problem of convolution operation, this paper uses discrete fast Fourier transform instead of convolution operation, transforming convolution operation to simple multiplication, so that the amount of calculation is reduced. Experimental data show that the improved algorithm has better color fidelity and can improve the computing speed compared with the existing algorithms, and the visual perception of color images has been improved.Under foggy weather conditions when processing images, it causes some problems such as the halo and whitening effects appeared in the region of bigger gradient, or the image contrast insufficient. In order to solve these problems, this paper proposes a Retinex defogging algorithm based on contrast limited adaptive histogram equalization(CLAHE). This algorithm uses guided filter instead of Gaussian filter in the original algorithm, compared with the original Gaussian filter, it has lower algorithm complexity and can preserve images' edge gradient, with the advantages of significantly improve the phenomenon of halo. The reflection component is processed by contrast limited adaptive histogram equalization. The experimental results show that this algorithm can apply to different fog-degraded images, can enhance the details of original images information, and achieving a good effect on both noise suppression and edge reservation.