A Variational Level Set SAR Image Segmentation Approach Based on Statistical Model

SAR image automatic segmentation is a hard work due to the presence of speckle noise. In this paper, a variational level set approach for SAR image is presented. A new energy functional is defined by taking account of a statistical model of speckle noise. The energy functional is with respect to level set function, which is obviously different from the energy functional with respect to parameterized curve in general level set approach. Segmentation is implemented by minimizing the energy formulation via level set approach. The performance of the approach is verified by MSTAR SAR images. It shows that the energy well describes the property of SAR image, thus accurately and automatically extracts the regions of interest in SAR image but without any speckle pre-processing step.

[1]  Carlos Vázquez,et al.  SAR image segmentation with active contours and level sets , 2004, 2004 International Conference on Image Processing, 2004. ICIP '04..

[2]  V. Caselles,et al.  A geometric model for active contours in image processing , 1993 .

[3]  Hong-Bo Yang,et al.  Nonparametric methods for image segmentation using active contour , 2004 .

[4]  Baba C. Vemuri,et al.  Shape Modeling with Front Propagation: A Level Set Approach , 1995, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  Guillermo Sapiro,et al.  Geodesic Active Contours , 1995, International Journal of Computer Vision.

[6]  Amar Mitiche,et al.  Multiregion level-set partitioning of synthetic aperture radar images , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7]  Frédéric Galland,et al.  Minimum description length synthetic aperture radar image segmentation , 2003, IEEE Trans. Image Process..

[8]  T. Chan,et al.  A Variational Level Set Approach to Multiphase Motion , 1996 .

[9]  Tony F. Chan,et al.  Active contours without edges , 2001, IEEE Trans. Image Process..

[10]  Chunming Li,et al.  Level set evolution without re-initialization: a new variational formulation , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).