Research On Algorithm Of Image Edge Detection Based On Wavelet Transformation

With the rapid development of digital image processing technology, new theories continue to emerge, such as the combination of wavelet analysis, fractal geometry, mathematical morphology, Markov random field, etc. This makes the digital image processing have a considerable development and show a strong vitality.In this paper, the main line is edge detection, analysis of several classical edge detection operator, and the comparison of the characteristics of each operator, but these methods suppress noise is weak, due to noise, often result in wrong inspection or missed phenomenon. The wavelet transform has good time-frequency analysis characteristics and high precision positioning. These properties can be detected well with the edges of the image.So, this paper makes a deeply study of the wavelet-based image edge detection algorithms. First of all, it makes a research of the basic theory of wavelet transform, such as discrete wavelet transform, continuous wavelet transform, dyadic wavelet, multi-resolution analysis and Mallat algorithm; and introduces several wavelet-based image edge detection algorithms. Including edge detection based on lifting wavelet algorithm, wavelet packet transform and edge detection, based on the B-spline wavelet transform fast multi-scale edge detection algorithm, and on this basis, it puts forward an improved B-spline wavelet-based edge detection algorithm.