A Weak Edge Detection Algorithm Based On Nonlinear Transform Of Gray Levels

In computer vision and image processing, edge detection refers to the process of localizing edge points in an image. It preserves useful structural information about object boundaries in an image and reduces the amount of data to be processed, so it can simplify the vision processing and image analysis tasks. Therefore, edge detection is a fundamental topic in computer vision and image processing in the last several decades and of which the weak edge detection is a problem that hasn't been solved very well. In view of enhancing the ability to detect weak edge, a weak edge detection algorithm based on nonlinear transform of gray levels is proposed in this thesis.In this thesis the background, current state and existing problems of weak edge detection are first described. Then four classical edge detection methods are discussed, including Canny's method, Meer's method and two adaptive methods. The defect of using these methods to detect weak edge is analyzed. Finally, a weak edge detection algorithm based on nonlinear transform of gray levels is proposed. It first transforms the image grays using a nonlinear function to enhance the contrast at the weak edge, it then makes a smoothing, calculates the gradient magnitude and makes a non-maximal suppression to get candidate edge points with one pixel width, finally it chooses a proper threshold using the gradient histogram to label edge points. Experimental results show that the proposed algorithm can detect weak edges effectively.