Pan-Tilt-Zoom Camera Networks.

摘要: Abstract Pan-tilt-zoom (PTZ) camera networks play an important role in surveillance systems. They can direct attention to interesting events the scene. One method achieve such behavior is a process known as sensor slaving: master (or more) monitors wide area and tracks moving targets provide positional information one slave more). The thus foveate at high resolution. In this chapter we consider problem of estimating online time-variant transformation human's foot position image fixed relative his head PTZ camera. achieves high-resolution images by steering detected view. Assuming planar scene modeling humans vertical segments, present development uncalibrated framework that does not require any 3D location be specified takes into account both zooming target uncertainties. Results show good performances localizing target's view, degrading when zoom factor causes lack feature points. A cooperative tracking approach exploiting instance proposed presented.