中文摘要：

基于贝叶斯框架下的稀疏重构方法，由于考虑了稀疏信号的先验信息以及测量过程中的加性噪声，因而能够更好地重建目标系数，然而传统的稀疏贝叶斯学习（SBL）算法参数多，时效性差。该文考虑一种新的稀疏贝叶斯学习方法方差成分扩张压缩（ExCoV），其不同于SBL中赋予所有的信号元素各自的方差分量参数，ExCoV方法仅仅赋予有重要意义的信号元素不同的方差分量，并拥有比SBL方法更少的参数。基于计算机层析成像技术框架下的ISAR成像模型，该文将ExCoV方法结合压缩感知（CS）理论将其进行ISAR成像，并从适用性和成像效果等方面与常用的极坐标格式算法（PFA），卷积逆投影算法（CBPA）和传统的稀疏重构算法进行比较，点目标仿真结果表明基于ExCoV的方法得到的ISAR像具有低旁瓣，高分辨率的特点，真实数据的成像结果表明该方法是一种比SBL更有效的ISAR成像算法。

英文摘要：

By taking into account of the prior information of the sparse signal and the additive noise encountered in the measurement process, the sparse recover algorithm under the Bayesian framework can reconstruct the coefficient better. However, the traditional Sparse Bayesian Learning （SBL） algorithm holds many parameters and its timeliness is poor. In this paper, a new sparse Bayesian learning algorithm named Expansion-Compression Variance-component based method （ExCoV） is considered, which only endows a different variance-component to the significant signal elements. Unlikely, the SBL has a distinct variance component on the all signal elements. In addition, the ExCoV has much less parameters than the SBL. Combined with the Compress Sensing （CS） theory, the ExCoV is used in the ISAR imaging model under the Computerized Tomography （CT） frame, and its applicability and the imaging quality are compared with the Polar Format Algorithm （PFA）, Convolution Back Projection Algorithm （CBPA） and the traditional sparse recover algorithm. The point scatter simulation verifies that the Inverse SAR （ISAR） image obtained by the ExCoV has low sidelobe and high resolution, and is not sensitive to noise. The imaging results of real data show that the ExCoV has more sparse ISAR image, indicating that it is a more effective and potential ISAR imaging algorithm.