SAR image segmentation based on mixture context and wavelet hidden-class-label Markov random field

In order to suppress the effect of multiplicative speckle noise on Synthetic Aperture Radar (SAR) image segmentation, a new SAR image segmentation algorithm is proposed based on the mixture context and the wavelet hidden-class-label Markov Random Field (MRF). In our paper, a wavelet mixture heavy-tailed model is constructed, and the hidden-class-label MRF is extended to the wavelet domain to suppress the effect of speckle noise. The multiscale segmentation with overlapping window is presented here to segment the finest scale of stationary wavelet transform (SWT) domain, and the classical segmentation method is still utilized at the coarse scales of discrete wavelet transform (DWT) domain, moreover, a mixture context model is proposed to combine the two different segmentation methods. Finally, a new maximum a posteriori (MAP) classification is obtained. The experimental results demonstrate that our segmentation method outperforms several other segmentation methods.