Long-time integration is an effective method to improve target detection performance in a noisy background. However, when detecting high-speed and maneuvering targets by long-time integration, it is easy to encounter the across range unit and across Doppler unit effects, which deteriorate the detection performance of algorithms with low computational complexity, such as moving target detection and hybrid integration (HI). The generalized Radon–Fourier transform (GRFT) detector has proven to have the best detection performance. However, the GRFT has a high computational burden for an ergodic search in a multidimensional motion parameter space and thus is hardly employed in real engineering applications. In this paper, we propose subspace HI (SHI) to achieve a good balance between detection performance and computational complexity. SHI first divides the parameter space into several equidimensional subspaces and moves them to the center of the coordinate system. Then, HI is implemented on all the subspaces, and all the HI results are finally fused. Through parameter space division and subspace movement, the values of the parameters in the subspaces are reduced, which increases the subaperture length, i.e., the time that the target echoes stay in one range-Doppler unit. By increasing the subaperture length, SHI gradually improves the detection performance with an increase in the computational complexity. Conversely, by shortening the subaperture length, the computational complexity of SHI can be reduced at the expense of the detection performance. Compared with HI and the GRFT, SHI achieves a better compromise between detection performance and computational complexity.

中文翻译：

---

一种用于高速机动目标检测的子空间混合积分方法

长时间积分是提高噪声背景下目标检测性能的有效方法。然而，在长时间积分检测高速机动目标时，容易遇到跨距单元和跨多普勒单元效应，从而降低了计算复杂度低的算法的检测性能，如运动目标检测和混合检测。集成（HI）。广义 Radon-Fourier 变换 (GRFT) 检测器已被证明具有最佳检测性能。然而，GRFT 在多维运动参数空间中进行遍历搜索具有很高的计算负担，因此很难在实际工程应用中使用。在本文中，我们提出了子空间 HI（SHI）来实现检测性能和计算复杂度之间的良好平衡。SHI 首先将参数空间划分为若干等维子空间，并将它们移动到坐标系的中心。然后对所有子空间进行HI，最后融合所有的HI结果。通过参数空间划分和子空间移动，减小子空间中的参数值，从而增加子孔径长度，即目标回波在一个距离多普勒单元内停留的时间。通过增加子孔径长度，SHI随着计算复杂度的增加逐渐提高检测性能。相反，通过缩短子孔径长度，可以以牺牲检测性能为代价降低 SHI 的计算复杂度。与 HI 和 GRFT 相比，SHI 在检测性能和计算复杂度之间实现了更好的折衷。