Multi-resolution Edge Detection Based on Alpha-stable Model in SAR Images Using Translation-Invariance Contourlet Transform

In this paper we present the enhanced translation-invariance contourlet transform based on alpha-stable model and its application in edge detection of synthetic aperture radar (SAR) images. The translation-invariance contourlet transform is built upon the translation-invariance wavelet decomposition and nonsubsampled directional filter banks (DFB). Due to the impulsive nature of multi-revolution coefficients achieved, by the contourlet transform, they can be accurately modeled by alpha-stable model. Based on alpha stable statistic model of multi-resolution coefficients, maximum a posteriori (MAP) estimator can enhance them. Relating and fusing enhanced coefficients at different resolutions can take account of noise suppression, the integrity of edges and the veracity of position so as to improve probability of detection reducing probability of false alarm. Experimental results show that the proposed algorithm achieves better performance than the statistical edge detector of SAR image