Variational Segmentation of Polarimetric SAR Image By Using a Continous Potts Model and Automatic Initialization
In this chapter, a polarimetric synthetic aperture radar (PolSAR) image segmentation approach is proposed in the variational framework. A novel continuous Potts model which takes advantage of the complex Wishart distribution is built for PolSAR image segmentation. Moreover, an automatic initialization technique is adopted to initialize the segmentation process. The automatic initialization approach can determine the number of clusters by PolSAR data itself. Compared with previous variational PolSAR segmentation approaches, the proposed approach makes use of the scattering characteristic and statistical characteristic together to segment PolSAR images in a completely unsupervised way. Experimental results demonstrate the effectiveness of the proposed approach. Without any artificial supervision, the proposed approach can produce superior segmentation results than the traditional level set method and \( Wishart - H - \alpha \) classification approach.
KeywordsVariational PolSAR image segmentation Level set Automatic initialization
This work is supported by the National Natural Science Foundation of China under Projects 60802065 and the Fundamental Research Funds for the Central Universities under Projects ZYGX2009Z005.
- 4.Kass M, Witkin A, D (1987) Snakes: active contour models. Int J Comput Vision:321–331Google Scholar
- 12.Fang Cao, Wen Hong, Yirong Wu, Pottier E (2007) An unsupervised segmentation with an adaptive number of clusters using the SPAN/H/α/A space and the complex wishart clustering for fully polarimetric SAR data analysis. IEEE Trans on Geosci Remote Sens 45(11):3454–3467Google Scholar
- 13.Chunming Li, Chen yang Xu, Chang feng Gui, Fox MD (2005) Level set evolution without re-initialization: a new variational formulation. In: Proceedings of the 2005 I.E. computer society conference on computer vision and pattern recognition (CVPR’05), San Diego, CA, United States 1063–6919/05Google Scholar