Wavelet analysis of a moving object trajectory
Cuihong Qi, Hong Shu, Jun Pang, Diana Quesada Sedas · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2006
In this paper, we make wavelet analysis of the moving object trajectory to discover some motion patterns locally and globally. Particularly, the trajectory is represented with a sequence of spatial coordinates over time. Boundary extension is made previously for meeting the mathematical preconditions of wavelet transformations. The spatial point, composed of coordinates x and y, is regarded as a function of time. After a wavelet transformation of the moving object trajectory, the wavelet coefficients of modulus maxima are selected for identifying sudden speed-changing points. Further, we make a cubic spline interpolation of pseudo-modulus maxima wavelet coefficients for reducing the error of the reconstructed motion trajectory. Some experimental results have shown that motion maintenance and trajectory generalization are achieved through wavelet analysis of the moving object trajectory.