Camera calibration based on vertical objects

In intelligent surveillance, how to calibrate camera in absence of calibrated objects or scene´s structure information or camera´s motion information, is a significant topic. This paper proposes a novel approach based on vertical objects using 4~5 positions labeled by human in video sequences with knowing height and obtains vertical vanishing point and plane vanishing line. Considering the importance of the vertical vanishing point, the plane vanishing line and the sensitivity of those two to noise, an optimal algorithm to vanishing point and line is proposed; Then according to the properties of vanishing point and vanishing line being function of camera parameters, all the intrinsic and extrinsic parameters are solved; Finally, this paper shows the optimal number and position of human in camera field. Both computer simulation and real data validate robustness and effectiveness of the algorithm.