Research On Improvement Of Color Image Graying Algorithm

With the development of computer technology and people's increasing dependence on computers,digital media technology has begun to fill our lives,and color images have become the main carrier of digital technology with its rich visual effects,and it has become the work of many researchers.Focus.Despite this,the grayscale image does not fade out of our field of view,but instead has the advantage of displaying most of the features of the image with less information,active in medical imaging,black and white photography and printing,face recognition and target tracking.field.Therefore,the basic color image graying technology is still a hot spot in image research.The process of color image graying is essentially a process of down-converting a three-dimensional color image into a one-dimensional gray image,and information loss may be avoided in this process.Therefore,how to maintain the visual information of the original image to the maximum extent in the grayscale process and make the result conform to human visual habits is the focus of this thesis.In this thesis,two kinds of matching gray-scale algorithms are proposed for the night scene image,foggy image and cold-toned image.Finally,the experimental results of this thesis have achieved good results under both qualitative and quantitative evaluation methods.The specific research work of this thesis is as follows:(1)A grayscale algorithm for the night scene image and the foggy image that is significantly maintained is proposed.To change the method of automatic grayscale mapping of the whole image by the previous grayscale algorithm,this thesis uses the frequency-tuned saliency detection technology to obtain the saliency map of the original color image for the characteristics of night scene image and foggy image.The target area of interest is found,and then the guide map is processed,and the polynomial parameter model is constructed by using the saliency value of the pixel in the guide map and the weight of the RGB three-channel in the original color image.Constructing an energy optimization function,using the guide map to change the reference contrast value in the energy optimization function,enhancing the contrast of the target region in the finally obtained grayscale result,and finally obtaining the grayscale result of the significantly maintained night scene image and the foggy image..(2)A grayscale algorithm for cold tones images is proposed.In the past,the gray-scale algorithm only considers the inherent information such as the contrast of the image without considering the shortcomings of the color composition of the image.In this thesis,based on the color composition of the cool-toned image,a cold-tone image gray based on Lu et al.Algorithm.After expounding the basic principle of Lu et al.algorithm,the algorithm of Lu et al.is cumbersome and easy to lose information in the process of color space conversion.The RGB color space which is more suitable for processing cold-tone images is selected.The mapping function of the color space finally results in a grayscale result of the cold image with better effect.