Research On Wavelet Multiscale Image Edge Enhancement Method And Its Application

Edge detection is an important task in image processing.It is the main tool for pattern recognition,image segmentation and scene analysis.The edge detection method is basically a high-pass filter that can be used to extract edge points in an image.This subject has attracted many researchers and achieved many results.J.Canny proved that the optimal filter can be approximated by the derivative of Gauss function.Some classical edge detectors can not deal with noisy images well,because their signal-to-noise ratio is not the largest through the detection filter.S.Mallat proposed an image denoising method based on wavelet modulus maxima.If we assume the existence of white noise,we will produce singularities with negative Lipschitz regularity.Therefore,by observing the amplitude of modulus maxima at different scales,we can preserve the edges propagating to a certain rough scale and remove other edges dominated by noise.In this article,we will explain the mechanism of the edge detection method from the perspective of wavelet multi-scale,and propose a new method of constructing an edge detection enhancement filter using wavelet transform.First,we briefly introduced the development of wavelet analysis,and then reviewd some of the main traditional edge detection methods and explained them with continuous dyadic wavelet transform.Traditional edge detection methods are good at processing high-quality images,but they are usually not obvious enough for noisy images because they cannot distinguish edges of different meanings.Secondly,it is explained by dyadic wavelet transform,and then the continuous edge detection on multi-scale of wavelet is discussed,and a cascade algorithm of dyadic wavelet transform is proposed.The dyadic wavelet transform-based edge enhancement algorithm proposed in this paper combines wavelet transform coefficients on a series of scales,which significantly improves the results.Finally,a continuous algorithm is proposed to realize the wavelet multi-scale image edge enhancement algorithm.