Contents Based Color Image Edge Detection By Using The Centerline Extraction Method

The edge information is the most significant information above all in an image.Edge detection is a fundamental and the first step in image processing, image analysisand understanding and many computer vision applications. Thus, obtained the edgeinformation of an image can generally got the content information of the image.Edgeinformation is the most intuitive information to distinguish one thing fromanother.The edges of the gray image contains less information than in the color image,The edges of the color image contain nearly70%to80%information of an image, Sothe color image edge detection is particularly important for obtaining the imagecontent.At present, although there have been many sophisticated edge detectionmethods，but mostly methods unable to keep a balance between accuracy andcompleteness of the testing results. The adaptability of these methods is bad, so thatcan’t get a good application. Looking for a suitable color image edge detectionalgorithm became the key of image understanding and processing.This paper mainlystudies the center-path extraction method for the edge information in an image. Edgedetection can get the edges of outline and the detail texture of the object at the sametime, while too much detail edges could interfere with the discrimination. Therefore,in view of extracting the center path of the edge information and filting out theinterference of texture edges,we carried out in-depth study.This article first briefly introduced the research status and the developmenthistory of the edge detection method, made the analysis and research of theenhancement and edge extracting method. And then made a detailed analysis andresearch about the hessian matrix based center line extraction method, and putforward the color image edge detection method based on the hessian matrix，thenmade design, testing, simulation and analysis this algorithm, Experimental resultshave confirmed that the proposed detection result is more continuous and clear, alsocontaining more details. Good balance between the integrity and accuracy of edgedetection.