A Noncentral and Non-Gaussian Probability Model for SAR Data

被引:0
|
作者
Cristea, Anca [1 ]
Doulgeris, Anthony P. [1 ]
Eltoft, Torbjorn [1 ]
机构
[1] Univ Tromso, Dept Phys & Technol, Earth Observat Lab, N-9037 Tromso, Norway
来源
关键词
Speckle; Compound statistical model; Inverse Gaussian; MAP filter;
D O I
10.1007/978-3-319-59129-2_14
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
A general compound statistical model for coherent imaging is developed and tested on single-channel Synthetic Aperture Radar (SAR) data. In this formulation, coherent scattering is taken into consideration and the texture is modeled using an Inverse Gaussian distribution. Parameter estimation is conducted via an Expectation Maximization (EM) scheme. A Maximum a Posteriori (MAP) speckle filter based on this model is also implemented. The filter shows good smoothing capabilities and preserves details in the selected scene, showing promise for target-detection applications.
引用
收藏
页码:159 / 168
页数:10
相关论文
共 50 条