A Gaussian Mixture PHD Filter Based on Energy Accumulation

Aiming at the problem that tracking performance of the Gaussian mixture probability hypothesis density (GM-PHD) filter is inaccurate under low signal-to-noise ratio (SNR), a GM-PHD filter based on energy accumulation of trajectory is proposed. Firstly, several constant frames are selected to be the input of the pipeline filter, remaining possible straight lines within the speed limit. Then, accumulate the pixel value of the possible straight paths as their energies to further reduce false alarms. Finally, the number and status of targets are obtained through the iterative update of the classic PHD filter. The simulation results show that the proposed method achieves a favourable tracking performance, especially in the accuracy of the target number and target state estimates.