Multi-Hough Transform Track Initiation for Detecting Target with Constant Acceleration

Track initiation is an important part in multi-target tracking, especially for the lower observable targets under the condition of heavy clutter. Multi-Hough transform track initiation for detecting target with constant acceleration in heavy clutter environment is proposed. In this algorithm, standard Hough transform is used to filter clutter first. Then sift the candidate tracks by using randomized Hough transform and minimum distance method. Finally, the true track number and parameters are given by subtractive clustering. Simulation results show that this algorithm can keep higher true track initiation probability but lower false track initiation probability. By means of multi-Hough transform and subtractive clustering, the heavy clutter is filtered, the probability of sifting true track from the candidate tracks is increased, and the time for track initiation is shorten. This method show better performance than other methods.