传统恒虚警(CFAR)检测算法采用滑窗结构进行动目标检测,但在实际应用中,该结构的冗余操作会占据大量有限的雷达硬件资源。为解决上述问题,文中提出一种新的改进二维CFAR检测算法。该算法使用快速卷积结构取代传统CFAR算法的滑窗结构,从而可利用频域特性对其进行快速求解;同时,考虑常规CFAR在面对邻近多目标情况时检测性能急剧下降的问题,在卷积结构的基础上设计新的卷积核用于改善判决能力。实验结果表明:文中快速二维CFAR算法相比传统CFAR算法在运算量上有了极大的降低,还可以有效检测邻近目标,算法也易于在工程中实现。

Sliding-window structure is used in traditional constant false alarm rate (CFAR) algorithms to detect moving targets, however in application, the redundant operations of this structure will occupy large radar hardware resource. Based on above problem, a new fast two-dimensional CFAR algorithm is proposed in this paper, the sliding-window structure is replaced by fast convolution structure and thus the property of frequency domain can be utilized to solve the problem. Then considering the degrading detetion ability for neighbor multi-targets of conventional CFAR, a new convolution kernel is designed to improve the ability of distinguishing neighbor targets. The numerical experiment results show that the proposed algorithm has a significant saving in calculation than conventional CFAR, it can also detect neighbor moving targets effectively. Furthermore, this algorithm is easy to achieve in engineering.

TN957.51