Robust Edge Detection Based on Anisotropic Mathematical Morphology and Scale Multiplication in NSCT Domain
A novel edge detection method based on anisotropic mathematical morphology and scale multiplication in nonsubsampled contourlet transform (NSCT) domain is proposed to obtain a superior and robust performance under heavy noise. One preliminary result is obtained using anisotropic morphological gradient of the low-frequency component, yielding a single-pixel response with few pseudo edges. Due to the great ability of NSCT to localize distributed discontinuities such as edges, scale multiplication results of high-frequency components can get rid of a large amount of noise and produce well-localized edge candidates. The final result is a fusion of the detection results of low-frequency component and high-frequency components. Detailed experiments compared with other state-of-the-art methods demonstrate that the proposed method has a superior performance of edge detection and is quite robust even under heavy noise.