Edge Detection Based on Fusion of Wavelet Transform and Mathematical Morphology

The key issue of edge detection is to filter more noises while obtaining more edge details. This paper proposed a novel edge detection method based on the fusion of wavelet transform and mathematical morphology. Firstly, decompose image into high and low frequency parts via wavelet transform. Secondly, process the high and low frequency parts respectively: apply multi-scale wavelet transform to detect edge in high frequency part; and apply mathematical morphology to detect edge in low frequency part. Finally, merge the edge images using multi-scale edge fusion and extract the edge from merged edge image. Experimental results show that the proposed method has great effects in noise suppression and edge detail preserving compared with that of wavelet transform based or mathematical morphology based edge detection method.