摘要
Alpha-stable distribution is utilized to model statistical property of synthetic aperture radar (SAR) image, but it is still a challenge to accurately estimate their parameters due to the lack of analytical expression for its probability density function. A new estimator called pseudo simulated annealing (PSA) is introduced. Simulation results show the PSA estimator outperforms three other estimators that we tested. Finally, a classification algorithm is proposed for SAR image based on Alpha-stable distribution in Markov random field (MRF) framework. Experimental results over TerraSAR-X data demonstrate the efficiency of Alpha-stable distribution in SAR image statistical modelling and classification.
- 出版日期2013
- 单位武汉大学