Conic and circular points based camera linear calibration

This paper proposes an approach for camera linear calibration based on conic and circular points. The calibration template which contains a circle and its any 3 diameters is used and taken three or more pictures from different orientations. Conic is fitted with image feature points, coordinates of vanishing points on diameters and coordinates of tangents through the endpoints are calculated by using the projective invariance and the harmonic conjugates in projective geometry. By the orthogonal relations between diameters and tangents and the corollary of Laguerre Theorem, coordinates of circular points are solved. Using the nature of circular points establishes constraint equations on the intrinsic parameters, and all camera intrinsic parameters are linearly solved. This approach has simple computational process, without knowing the metric information of the circle. Simulation and real experiment are given, and the results show that the approach has the higher accuracy and strong robustness. © 2010 Binary Information Press.