Hough Transform Track Initiation Algorithm Study

Multi-target track initiation in low SNR and low SCR is the basic problem in target tracking. Hough transform has the low sensitivity to local fault, high robustness against stochastic noise and good adaptability in the parallel and real-time processing. Hence Hough transform track initiation method is studied in this paper and the main contributions are as follows:1. The important track initiation methods are summarized in this paper. The principle and development of Hough transform are introduced, and the problems of Hough transform based track initiation methods are also proposed.2. We introduce and further supply simulation-based analysis of five classical Hough transform track initiation methods, i.e. standard Hough transform, randomized Hough transform, modified Hough transform, sequential Hough transform and subarea Hough transform.3. A fast randomized Hough transform based track initiation method is proposed for the drawback of large computation of randomized Hough transform. This method uses the movement information of target and the termination rule to reduce invalid samplings. Simulation results show the performance of sampling is improved and the time of track initiation is obviously reduced.4. A Multi-Scale Clustering based Hough transform track initiation method (MSC-HT) is proposed for track clustering in the traditional Hough transform based track initiation method. The proposed method first obtains the rough candidate tracks via Hough transform with the lower threshold. Then the Multi-Scale Clustering method is applied to refine the candidate tracks.5. Three special methods which are modified Hough transform, randomized Hough transform and two-hierarchical Hough transform are applied in OTHR environment.