论文部分内容阅读
高斯粒子概率假设密度(PHD)滤波往往假定杂波密度参数已知,这种做法对于实际应用是不现实的。此外,杂波的参数值通常依赖于环境条件,可能随时间发生变化。因此,多目标跟踪算法中需要实时准确估计杂波密度的参数。基于此,提出了一种多目标跟踪的区域杂波估计方法。首先根据量测信息在线估计出场景中的杂波数目,然后估计落入目标附近感兴趣区域的杂波数,并估计每个目标感兴趣区域杂波强度。仿真结果表明,在复杂场景下算法的跟踪性能明显优于未进行杂波估计的多目标跟踪算法,提高了跟踪的实时性和跟踪精度。
Gaussian particle probability hypothesis density (PHD) filtering often assumes that the clutter density parameters are known, which is not practical for practical applications. In addition, clutter parameter values are usually dependent on environmental conditions and may change over time. Therefore, the multi-target tracking algorithm needs to accurately estimate the parameters of the clutter density in real time. Based on this, a multi-target tracking clutter estimation method is proposed. Firstly, the number of clutter in the scene is estimated online based on the measurement information. Then the number of clutter that falls into the region of interest near the target is estimated and the clutter intensity of each region of interest is estimated. The simulation results show that the tracking performance of the algorithm is better than that of the non-clutter-based multi-target tracking algorithm in complex scenarios, which improves the real-time tracking and tracking accuracy.