A New Adaptive Zoom Algorithm for Tracking Targets Using Pan-Tilt-Zoom Cameras

We present a new method to adaptively configure the focal-length of a pan-tilt-zoom camera to track a point target with improved accuracy. The target tracker is implemented by using a Rao-Blackwellized particle filter. The focal-length of the camera is configured to include a time varying percentage of the projected particles on the camera image plane. The performance of the adaptive zoom algorithm was evaluated using Monte Carlo simulations and found to significantly reduce tracking errors compared to a previous method presented by the same authors

[1]  Fredrik Gustafsson,et al.  Sequential Monte Carlo filtering techniques applied to integrated navigation systems , 2001, Proceedings of the 2001 American Control Conference. (Cat. No.01CH37148).

[2]  Rafael C. González,et al.  Local Determination of a Moving Contrast Edge , 1985, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  D. Morrell,et al.  Adaptive foveal sensor for target tracking , 2002, Conference Record of the Thirty-Sixth Asilomar Conference on Signals, Systems and Computers, 2002..

[4]  L. Li,et al.  Target tracking with an attentive foveal sensor , 2000, Conference Record of the Thirty-Fourth Asilomar Conference on Signals, Systems and Computers (Cat. No.00CH37154).

[5]  Stan Birchheld,et al.  An Introduction to Projective Geometry for Computer Vision , 2022 .

[6]  Himanshu Shah,et al.  An adaptive zoom algorithm for tracking targets using pan-tilt-zoom cameras , 2004, 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[7]  D. Morrell,et al.  Target tracking and data fusion using multiple adaptive Foveal sensors , 2003, Sixth International Conference of Information Fusion, 2003. Proceedings of the.

[8]  Darryl Morrell,et al.  The use of particle filtering with the unscented transform to schedule sensors multiple steps ahead , 2004, 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing.