Computing camera parameters using vanishing-line information from a rectangular parallelepiped

The normalised image of the absolute conic and its application for zooming camera calibration

“…This choice has been mathematically justified in [27], and we can observe that in the case of an aspect ratio close to one the normalisation introduced is very close to the denominator in Eq. (14). This means that the system is well conditioned and approximates very closely the more complicated geometric distance with the advantage of not requiring non-linear minimisation.…”

“…Previous works, in the case of non-zooming cameras, have taken advantage of invariants to decouple the camera parameters into simpler sub-problems and thus guarantee that the number of unknowns of each sub-problem is constant. Examples include Vanishing Points (VPs) [13][14][15][16][17][18][19][20][21], which are invariant to translation, and the Image of the Absolute Conic (IAC) [22,20,[23][24][25][26][27][28], which is invariant to translation and rotation. In this paper, the invariance properties of the IAC are extended to zooming, by defining the Normalised Image of the Absolute Conic (NIAC), which characterises uniquely the camera parameters independent of position, orientation and zooming.…”

Using Points at Infinity for Parameter Decoupling in Camera Calibration

“…Our approach presents some similarities with previous methods based on vanishing points [4], [6]- [9], [11], [20], [31], [32]. VPs have strong invariance properties, however, there are usually not many in images and they are difficult to compute.…”

“…Namely, we do not have a strict correspondence, but only a constraint that establishes that a VP should lie on the image line. This is fundamentally different to other camera calibration methods that propose computing the VPs from parallel lines before estimating the camera parameters [4], [6]- [9], [11], [20], [31], [32]. It should also be noted that the equations defined in 1) are similar to the estimation of the fundamental matrix via the 8-point algorithm [15].…”

Camera Calibration using Vanishing Points

“…The underlying theme in the calibration processing described here is the use of vanishing point and vanishing line information. Related methods are [3], [8], [2]. …”