The Grayscale Image Colorization Algorithm Research Based On Video

The color gray-scale image is an important component of the computer imageprocessing, and computer experts concern this hot topic in recent years. It’s mainlypurpose is that the color information of the color image moves to the grayscale image,so that the gray-scale image also has the color information. At present, satelliteimages, infrared images, medical images, most of them are still gray images or singleimages. Because the human eye’s ability to distinguish a gray-scale image is worsethan a color image, at the same time a color image contains more color information.Therefore a grayscale image colorization is a problem to be solved, and it can bringtremendous business value in the economic.This article describes the current commonly used the color space in the field ofthe computer image, and analyzes the advantages and disadvantages of these colorspaces and scope of the use. Focused on the αβ color space, the color space is used inthe grayscale image colorization process. Then analyzes the algorithm, which wasproposed by Welsh, it can be used in grayscale image colorization. It was called byWelsh algorithm. The study show that Welsh algorithm has a problem that it is tooslow in search for the best match of pixel values in the color image. So this articleproposes an improved search strategy in order to solve this problem. The method isthat take the interval access point approach in the color image, reducing the number ofcycles. At the same time, there is part of the noise in the image after grayscale imagecolorization by the improved Welsh algorithm. Therefore, in order to remove noiseinterference, using the guidance filtering processing means. At last,the classic paperimages and the actual image are choosed. The subjective appraisal and objectiveappraisal are used in using the guidance filtering images and without using theguidance filtering images. Experiments show that guidance filtering in the grayscaleimage colorization process can remove most of the noises. So that the final resultimage becomes more clear and natural.For the grayscale video processing, using the vision persistence of the humanphenomenon. First, each of the seconds, video image is divided into30frames still image, so groups of gray-scale image are obtained. Then using the grayscale imagecolorization algorithm in the article, coloring each of the gray images, a group ofcolor images are obtained. Then using the phenomenon of the vision persistence, theimages will display30frames per second. This time the human eye will feel that a setof video sequences, which can obtain the color of the video.