Algorithm Research Of Edge Detection To Digital Image Based On Grey System Theory

Along with the rapid development in science and technology,the creation of some new theories such as Wavelet Analysis,Markovian Random Field,Grey System Theory made the image processing technics develop considerably.This thesis did a research on how to apply the Grey System Theory to the field of digital image filter and edge detection aimed at exploring the appliance and feasibility of the Grey System Theory in image processing fields.It has made a solid foundation for the further research with very important theoretical significance and practical value.new theories and new ideas.This thesis firstly carried on the thorough analysis to the grey theory about its purpose,significance and the late researching state in image processing domian. Secondly,in the charpter 2,this paper summerized and generalized the grey theory's researching content,main character,basic theories and the application in image processing.Thirdly,in the charpter 3 and 4,the paper elaborates the concrete application in image filter and image edge-detection.It made a comparison between the new algorithm and the traditional algorithm.The paper objectly estimated the new algorithm's superiority and deficiency from many angels.Lastly,it put forward some improvable ideas and the further researching directions.This paper's research achievements and innovation are as follows:The articl exploratively applied grey correlative degree in both image denoise and image edge-detection at the same peper:①Aimed at resolving the problems of the median filter algorithm in enhancing the PSNR with limits and remaining the details of a picture imperfectly,a two-step denoising method with double threshold based on grey correlative degree is presented.The experimental result shows that this proposed algorithm could effectively improve PSNR and subjective effects of a picture under the conditions of high noise.The experiment result indicated that,the new algorithm can successfully differentiate noise_point and edge_point with the standard based on grey correlative degree.Even on the condition of higher noise ratio this algorithm can also make a acurate assignment about grayscale to the center of the filter_window protecting the edge and details of the image with a thorough removal to the noise_point;②Aimed at deficiency of the traditional edge-detection,a new algorithm based on grey correlative degree is presented.The experiment result indicated that, the new algorithm has a acurate recognition to the edge_point especially to the weak edge.The new algorithm can precisely locate the edge and form continuous outline on the condition of the best threshold.It achieved an aparent predominance in the experimental detective comparison among all those algorithms.