Research Of Image Edge Detection Algorithm

Image edge detection and extraction play a very important role in image processing,and have high value in applications. How to improve the accuracy of edge detection hasbeen the difficulty of the edge detection, it is also the focus of many scholars in research.Chapter1in this thesis introduces the research background and significance of digitalimage processing and edge detection techniques, and analyzes the problems in this field.Chapter2introduces the edge detection theory and some basic algorithms of thegrayscale image, and tells the advantages and disadvantages of these algorithms byanalyzing the relative experimental results.Chapter3, for avoiding the shortcomings of the Kirsch edge detection operator,presents a multi-threshold selection and edge connection method. Firstly, this methoddenoises the original image with Gaussian filter and use the traditional Kirsch operator tocalculate the gradient; then set a lower threshold and adaptive threshold to protect the weakimage edges and extract the real edges; finally connects the edges by use of the improvededge linking algorithm. The experimental results show that this method has the merits ofmaintaining weak edges and good connectivity.Chapter4first introduces two basic color space and two types of basic color imageedge detection methods, then defines the color values of the color image based on theCIELab color space, and lastly uses the modified edge connection method to detect thecolor image edge with the method introduced in Chapter3and obtains good edge detectionresults.