Self-Calibration of Stationary Cameras

  • Authors:
  • Richard I. Hartley

  • Affiliations:
  • G.E. CRD, P.O. Box 8, Schenectady, NY, 12301/ E-mail: hartley@crd.ge.com

  • Venue:
  • International Journal of Computer Vision
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

A new practical method is given for the self-calibration of a camera. In this method, at least three images are taken from the same point inspace with different orientations of the camera and calibration iscomputed from an analysis of point matches between the images. Themethod requires no knowledge of the orientations of the camera.Calibration is based on the image correspondences only. This methoddiffers fundamentally from previous results by Maybank and Faugeras on self-calibration using the epipolar structure of image pairs. Inthe method of this paper, there is no epipolar structure since allimages are taken from the same point in space, and so Maybank andFaugeras‘s method does not apply. Since the images are all taken from the same point in space,determination of point matches is considerably easier than for imagestaken with a moving camera, since problems of occlusion or change ofaspect or illumination do not occur. A non-iterative calibrationalgorithm is given that works with any number of images. An iterative refinement method thatmay be used with noisy data is also described. The algorithm is implemented andvalidated on several sets of synthetic and real image data.