The Optimization Of Automatic Color Equalization Algorithm

In image processing, color equalization is a basic application, and so far lots of classic algorithms have been constantly researched to overcome the color defects, increase the saturation of images, improve the local or overall image contrast so as to achieve a better visual effect.Automatic color equalization (ACE) algorithm gets more attentions and applications in view of effectiveness in image color correction and. However, the standard algorithm is based on image pixel and processes the image by point to point, so the computation complexity of ACE is too high to be suitable for many application scenes. Therefore, optimization for ACE is necessary for extensive application.The main contents in this paper are as follows:(1) An improved algorithm in this thesis is conceived based on traditional ACE and local linear look up table (LLLUT) method. Calculations are simplified and the calculating speed is promoted.(2) A gradual local linear look up table (GLLLUT) technique is analyzed, verified and used for high-resolution images to achieve better details restoration. Thus, set optimization for ACE is secured。(3) According to the calculation independence of the ACE algorithm for each channel, the Multi-core computing environment is fully taken advantage of to realize parallel computation without affecting the correction effect.(4) The optimization algorithm is applied to image processing. It produces a relative ideal effect for the color correction of overexposure, underexposure and low contrast images. Experiment results and data analysis shows that the devised method has not only similar correction result as ACE algorithm, but also better speed performance than LLLUT method. The optimization algorithm has achieved good results in image processing.