Self-calibration of projective camera based on trajectory basis.

Highlights • A linear method for self-calibration based on trajectory basis is presented. • The trajectories of non-rigid are modelled by the predefined DCT basis vectors. • Self-calibration becomes a linear optimal problem, which improves the algorithm robustness. • Experiments show that the presented method can effectively calibrate the camera. Abstract To calibrate the projective camera, this paper presents a linear method for self-calibration of a projective camera based on the representing of deforming object by the trajectory basis. The trajectory basis can be modeled compactly in the domain of the Discrete Cosine Transform (DCT) basis vectors which can be predefined independent of the observed images, and the number of unknowns can be significantly reduced. At the same time, the camera self-calibration becomes a linear optimal problem, which can improve the robustness of the algorithm. The experiments with both simulate and real data show that the presented method can effectively calibrate the camera. [ABSTRACT FROM AUTHOR]