Image Edge Detection Algorithm Based On Improved Dyadic Wavelet Transform And Morphology

The edge of an image is the most basic part of an image,which contains most of the information of the image.Therefore,image edge detection is an important basis for image segmentation and recognition.At present,the major problems of image edge detection are the contradiction between edge sharpness,location accuracy and denoising ability.In view of the shortcomings of morphological and wavelet transform edge detection algorithm,the above two algorithms are improved and fused.Specific contents are as follows:Wavelet transform edge detection algorithm can not get good results in image denoising and edge location because of the complexity of the image itself.Taking dyadic wavelet transform as an example,the accuracy of edge detection is further improved because of its good properties such as translation invariance,but the edge of detection has the disadvantage of discontinuous details.In order to overcome the shortcomings of edge detection algorithm based on dyadic wavelet transform,the smoothness of dyadic wavelet is improved and the appropriate wavelet filter is selected to overcome the shortcomings of the algorithm.Because the smoothness of the wavelet function is affected by the order of the vanishing moment of the wavelet function,this paper will focus on improving the order of the vanishing moment of B-spline dyadic wavelet,gives a new lifting scheme and parameters,and realizes a dyadic wavelet filter with high-order vanishing moment,symmetry and compact support.At the same time,a B-spline dyadic wavelet edge detection algorithm based on lifting is proposed.The algorithm can reduce the interference of image noise,and the image edge details are continuous.In mathematical morphology edge detection algorithm,different structural elements and morphological operators affect the ability of image edge detection and denoising.In this paper,we select structural elements with different directions to detect edge according are the improved morphological operator.The experimental results show that the edge of the noised image is smooth and the noise is less.Because the dyadic wavelet transform edge detection algorithm only handlesthe low frequency component of the image,neglects the information of the high frequency component of the image such that the edge contour details are few.However,the morphological edge detection algorithm has good denoising effect and rich details.In order to improve the continuity and accuracy of edges and preserve image details,this paper combines the two algorithms to put forward a fusion algorithm of the dyadic wavelet and anti-noise morphological,and the experimental results show that the edges obtained by using the fusion algorithm are continuous,clear and less noisy,which is superior to the above two algorithms in objective evaluation.