Canny Edge Detection of Complex Image Based on Double Sparse Decomposition
In the edge detection of complex images,how to eliminate the influence of non-coherent factors has always been the focus and difficulty of research.To solve the above problems,a double sparse decomposition method is proposed to separate the high-frequency feature vectors with strong interference in the image data.This method uses the Nonsubsampled Contourlet transform to pre-decompose the image,and then performs K-Singular Value Decomposition dictionary learning on the high-frequency compo-nents,and uses the obtained learning dictionary to sparse the image.According to the dictionary atom activity corresponding to the sparse coefficient.The image is decomposed into high and low frequency parts.And the Canny edge detection algorithm is im-proved,using the double sparse method to decompose the complex image to obtain the low frequency part,and then the Canny edge detection is performed on the purer low frequency image.Simulation experiments show that the sparse decomposition efficiency of the double sparse method is higher,and the Canny edge detection result combined with the double sparse method is clearer and more complete.
sparse representationlearning dictionaryK-Singular Value Decomposition(K-SVD)contourlet transformedge detection
国家科技期刊平台
NSTL
万方数据
