This paper presents a Deformation Detection Particle Filter (DDPF) for tracking multiple manoeuvring targets, improving accuracy through dynamic modeling and updates based on target deformation. The DDPF is compared to a basic Sequential Importance Resampling Particle Filter (SIR-PF), with results indicating that DDPF more effectively tracks targets experiencing rotations or scaling. Various methods for pose estimation and data association in clutter-free environments are discussed, with experiments conducted using real airshow videos.