Research On Grey Image Edge Detection Based On Grey Correlation Analysis

Based on the theory that the edge can be detected by grey collection analysis,this paper studies the basic principle of traditional algorithm for grey correlation analysis,and improves the defects that anti-noise performance is poor and threshold selection is subjective in traditional algorithm.In this paper,at first the paper mainly introduces the background knowledge of edge detection,the theoretical basis and research progress of edge detection based on gray relational analysis.In the third chapter,this paper improves the defects of traditional algorithms.Firstly,the median filter is used to smooth the image and enhance the anti-noise performance of the algorithm.Secondly,an adaptive threshold equation is proposed based on the human visual characteristics,which is composed of the average gray value of 3×3 neighborhoods around the detecting pixel,and overcome the flaws of the traditional algorithm threshold set subjectively.The result of simulation analysis indicates that the more pseudo-edge extracted by the improved threshold is improved by processing the eight neighboring pixels around the edge pixels.Finally,the simulation results show that the improved algorithm has better performance in these aspects of anti-noise and positioning error.In the fourth chapter,the detection performance of improved algorithm and the classical differential operator are compared with the method of experiment simulation from the aspects of anti-noise performance,positioning error,linear connection degree and edge continuity.Experiments show that the improved algorithm is more complete than the classical algorithm.The edge is fine and the positioning accuracy is higher.For the salt and pepper noise image with high concentration,the improved algorithm has also obtained the unmatched advantage of the classical algorithm.Finally,the paper discusses the applicability of the improved algorithm to different types of images and the time performance of the algorithm in different gray scale images,the experimental results show that the algorithm is applicable.