Digital Image Edge Detection Algorithm

Edge is the most basic features of the image, which contains useful information for identifying the image, and for people to describe or identify the target and explain the images providing the important characteristic parameters. Edge identification and extraction for the entire image scene recognition and understanding is very important, and edge detection of objects in pattern recognition is an important part of the feature extraction. Image edge detection has a long research history, has been a hot and focus problem of the image research. The actual processing of the images are generally mixed with noise, how to eliminate pseudo-edges caused by noise, and at the same time to ensure the accuracy of positioning the edge detection needs to be addressed an important issue.This paper introduces the model of image edge classification, edge detection in two difficult problems, as well as concrete steps to achieve edge detection, and then introduced the now widely used edge detection methods such as Sobel, Roberts, and Canny algorithm, and simulation experiments of these algorithms in noise and noise-free cases, a comparative analysis of test results of each algorithm. The core idea of these algorithms is to assume that the edge points correspond to the original image gray-level gradient of the local extreme points. However, when the image contains noise, these algorithms are very sensitive to noise, and often the noise will be detected as edge points, while the real edge of the interference due to noise may also be undetected. Second, proposed a Prewitt edge detection algorithm based on the traditional improved edge detection algorithm, this algorithm is the core idea of the traditional Prewitt operator of the two directions templates, expanded to eight directions, and use the image gradient mode, the edge point of the relevant and binding on the edge of the image edge tracking, and gradually the exclusion of false edges, accurate positioning edges to exclude noise. Then introduce the general use of wavelet transform edge detection method. Wavelet analysis theory as a new time-frequency analysis tool in signal analysis and processing has been a very good application. In the wavelet transform edge detection focuses on the principle of edge detection and specific implementation steps, and has done a simulation experiment using wavelet algorithm.