Color Image Intelligent Repairing Technology

The paper proposes an intelligent method based on color printed image repairing is different from the traditional which is used subjective manual techniques to repair color image in digitization and automated repair. Automated repair in modern printing image is the need to quantify the industry standard with digital technology products, which is thus an industrial scale and completely fill the practice to develop intelligent prepress system. To ensure the printing products qualities of the various color images conform to the national technical standards, and meet the requirements of high-quality printing. This requires various experts on the same printed color image quality criteria for unified and standardized, then forming a standard sample.Research focus of this paper is on the different types of the original color image characteristics, using digital processing functions in computing software, such as C++and Matlab, to develop about compensation and color correction which can be applied to the printing and digital print production standard algorithms. Analysis of the original digital images while to get the best plate parameters and process parameters.Research by extracting different types of color image original standard features to build the best quality color image processing parameters evaluated mathematical model, which can meet the business for all types of color image restoration and adjustment.Different from image defect repair, we will be based on the minimum error analysis, hypothesis testing, statistical analysis and other theories to analyze the characteristic curve of printing color images, color histogram and other digital graphic curve. Prepress image restoration intelligent demand is made on the industry in technology innovation and adjusts the level. It is the practice of the theoretical basis that using image color characteristic analysis, error analysis and mathematical statistics theory to guide new technology and adapt to new environments.This paper presents a new three-dimensional histogram optimization algorithm which is based on a three-dimensional probability of equalization. In the RGB color space to redefine a more uniform probability density function (PDF) and a cumulative density (CDF). Using three-dimensional color histogram equalization method instead of a one-dimensional histogram based on the R, G, B three-color respectively equalization that do not consider R, G, B three-color gray-scale level of the correlation. Then transferring image to HSV space and forming a new color image with nonlinear substitution on brightness and saturation. It is used to get more than sixty sets of experimental results, using three experiments to contrast eighteen images.