Studies On Canny Edge Detection Algorithm Based On DCT

The edges are the most fundamental features for images, edge detection technology is also a hot topic for researchers. So far, various edge detection methods have been appeared, but these methods have not a good performance on denoising which is complex. Facing this issue, the paper compare the ideal of Canny operator, taking image into DCT domain, handling the DCT coefficients, which can reduce the noise validly. The improved algorithm of Canny operator is proposed. Its main content is as follow:(1) Introduces the concept and present situation of edge detection, and describes the basic theory in detail about the traditional operators. According to analysis, we give their merits and demerits.(2) Usually, it should be denoised before image edge detection. This paper describes several common denoising methods, and proposes a method based on transform domain—DCT coefficients correction.(3) Highlight analyzes the shortcoming of Canny edge detection, and points out the poor ability to against the complex noises. Therefore, this paper proposes a Canny edge detection algorithm based on DCT transform. This method improves the ability to against noise and reserves more detail lines.(4) Not only the subjection evaluation but also the objection evaluation, our method performance is better than the familiar Canny edge detection algorithm. So it is an available manner for edge detection based on DCT.