基于贝叶斯估计的高精度ISAR成像  

High precision ISAR imaging via Bayesian statistics

作　　者：徐刚[1] 包敏[1] 李亚超[1] 邢孟道[1] 

机构地区：[1]西安电子科技大学雷达信号处理国家重点实验室,陕西西安710071

出　　处：《系统工程与电子技术》2011年第11期2382-2388,共7页Systems Engineering and Electronics

基　　金：国家重点基础研究发展计划(973计划)(2010CB731903);国家自然科学基金(61001211);高等学校博士学科点专项科研基金(20090203120013);西安电子科技大学基本科研业务费(JY10000902014)资助课题

摘　　要：基于贝叶斯估计原理,提出了一种贝叶斯逆合成孔径雷达(inverse synthetic aperture radar,ISAR)成像算法。基于最大后验概率准则建立ISAR成像模型,利用回波数据进行统计参数估计,以实现ISAR成像的自适应表征,从而提高ISAR成像的精度。特别是运动误差相位估计和ISAR图像的重构通过求解最优化问题实现,而未考虑误差相位的具体形式,具有较高的鲁棒性。此外,本文方法在低信噪比(signal-to-noise ratio,SNR)条件下,可以取得良好的聚焦效果,具有较好的噪声抑制能力。最后,贝叶斯估计问题转换为最优化问题进行求解,利用快速傅里叶变换及其逆变换(fast Fourier transform/inversed fast Fourier transform,FFT/IFFT)和矩阵对应点乘(Hadamard乘积)操作,有效提高该算法的效率。基于实测数据的实验验证了本文算法的有效性。A novel framework of high precision Bayesian inverse synthetic aperture radar （ISAR） imaging based on Bayesian statistics is proposed. The model of ISAR imaging is constructed in Bayesian formalism and statistic parameters are estimated accurately with data-driven in process. Then high precision ISAR imaging could be achieved by realizing adaptive ISAR image representation. Specifically, the novelty of the algorithm lies in its high robustness： phase adjustment is accomplished by solving an optimization problem, regardless of the formation of phase errors. Besides, the algorithm takes high capability of de-noise. Hence, the well-focused image could be achieved in low signal-to-noise ratio （SNR）. Finally, high efficiency could be ensured with fast Fourier transform/inversed fast Fourier transform （FFT/IFFT） and matrix Hadamard multiplication operations by converting Bayesian statistics into optimization. The experimental results using measured data confirm the validation of the proposal.

关 键 词：逆合成孔径雷达 贝叶斯统计 自聚焦 鲁棒性 噪声抑制 

分 类 号：TN957[电子电信—信号与信息处理]