A Research Of Edge Detection Based On Gray Relational Degree And Texture Analysis

edge extraction and detection plays an important role in image processing, and the pros and cons of the algorithms developed will directly affect the performance of the system.The traditional edge detection algorithm has been greatly developed, However, there are still many problems, particularly the coordination of the edge detection accuracy and the anti-noise performance. How to reconcile the contradiction between the two, and seeking simple and easy algorithms to implete edge detection is a very meaningful topic to be studied. Because the edge of the target entity, the image texture and even the noise could become a meaningful edge, For different edge features,it is difficult to find a kind of universal edge detection algorithm. The various existing edge detection algorithms are designed for a particular type of edge features, there are also their own limitations and inadequacies at the same time they reached the detection effect, Therefore, the field of digital image edge detection is a subject which needs further improvement and development. Because of this, this paper has carried out some of the work in the following:1.Using the gray system theory for image edge detection, the gray relational analysis method in the gray system theory made up of the defects caused by adopting the mathematical statistical methods to do a systematic analysis, It is available for each sample,no matter what’s the number of the sample or whether the sample has laws, more over, the computation is small, it is very convenient and can’t appear the situation which is inconsistent between the quantitative results and the qualitative analysis.2.Using the improved LOG operator, This operator can extract the more comprehensive edge and position more accurately.3.Using image texture analysis, texture is a reflection of an object structure, texture analysis of objects in the image can acquire the important information and it is an important means for image segmentation, feature extraction and classification of identification.4.Combining the three points above, the paper proposed a new edge detection algorithm and showed the effection,strong adaptability by the simulation experiments.