多尺度聚类Hough变换航迹起始方法

Multi-scale Clustering-based Hough Transform Track Initiation Method

针对基于Hough变换航迹起始方法存在的航迹簇拥现象,提出了一种基于多尺度聚类的Hough变换(MSC-HT)航迹起始方法。该方法首先对量测进行Hough变换,利用较低的阈值得到备选航迹;然后对备选航迹进行多尺度聚类,通过变化尺度参数寻优,自适应确定起始航迹数目和航迹参数,仿真结果表明了算法的有效性。

The difficulty of successfully initializing the target track while avoiding multiple initialized tracks derived from the same target is unavoidable in the traditional Hough transform based track initiation method. Here we proposed Multi-Scale Clustering based Hough transform track initiation method （MSCHT）. The proposed method first obtains the rough candidate tracks via Hough transform with the lower threshold, so that the target tracks can be remained. Then the Multi-Scale Clustering method is applied to refine the candidate tracks, so that multiple tracks from the same target are combined. Simulation result of multiple target track initiation in dense clutters shows that the MSC-HT works well.