H-PMHT track-before-detect processing with DP-based track initiation and termination

Abstract

Histogram probabilistic multi-hypothesis tracker (H-PMHT), based on probabilistic multi-hypothesis tracker, is a track-before-detect processing approach to detect dim targets. For the problem that H-PMHT cannot initiate new tracks and terminate tracks of disappeared targets, the authors propose a new dynamic programming (DP)-based H-PMHT algorithm, which can locate new targets by dealing with a few frames of sensor images and confirm disappeared targets according to their energy accumulation values along the existing tracks. With this sort of track initiation and termination mechanism, H-PMHT can be directly applied to realistic environments. Simulation results show that the DP-based H-PMHT algorithm can rapidly locate new targets and initiate tracks with a low false alarm rate, and quickly terminate tracks of disappeared targets with a low false termination probability.