基于改进正交匹配追踪算法的属性散射中心提取

属性散射中心模型是描述目标后向电磁散射特性的典型模型, 但其中传统的正交匹配追踪(orthogonal matching pursuit, OMP)算法提取模型时具有参数复杂度高、计算时间长等问题。对此提出一种基于稀疏字典的广义正交性的改进OMP算法, 快速定位模型位置参数值, 避免了正交匹配中的寻优过程, 从而降低算法的运算复杂度。通过对两类算法计算复杂度和计算精度进行多次蒙特卡罗实验比较得出，改进OMP算法提高了模型参数的估计精度与噪声鲁棒性, 且大幅降低了算法的运算复杂度, 相比于传统的OMP算法, 运算时间至少降低30%。

Attribute scattering center extraction based on improved orthogonal matching pursuit algorithm

The attribute scattering center model is a typical model to describe the backward electromagnetic scattering characteristics of the target, but the traditional orthogonal matching pursuit(OMP) algorithm has the problems of high complexity and long computation time when extracting the model parameters. Therefore, an improved OMP algorithm based on the generalized orthogonality of sparse dictionaries is proposed, which can quickly locate the position parameters of the model and avoid the optimization process in orthogonal matching for reducing the computational complexity of the algorithm. By comparing the computational complexity and computational accuracy of the two types of algorithms using multiple Monte Carlo experiments, the results show that the improved OMP algorithm improves the estimation accuracy and noise robustness of the model parameters, and significantly reduces the computational complexity of the algorithm. Compared with the traditional OMP algorithm, the computational time is reduced by at least 30%.