Camera Calibration with a Single Image

Objectives:

Camera parameters such as the focal length, the principle point and lens distortion can be estimated from the measurements of a single image when sufficient geometric object knowledge is available. The object knowledge used in the approach described here consists of parallelism and perpendicularity assumptions of straight object edges. In buildings parallel and perpendicular edges are usually abundant. Therefore, this method is often applicable for historic imagery of possibly demolished buildings taken with an unknown camera.

Approach:

Extraction of straight lines by a line-growing algorithm
Detection of three vanishing points using parallelism assumptions only (see the webpage on this subject)
Least-squares estimation of camera parameters using parallelism and perpendicularity conditions

The precision of the estimation depends on the availability of image lines of the three object orientations, the precision of the line measurements, and on the orientation of the image relative to the building. If some parameters cannot be estimated, they are excluded from the adjustment.

The procedure runs automatically, but depends on the correct automatic vanishing point detection. Although good results have been obtained (see the example below), large lens distortions could degrade tthe vanishing point detection. In that case, manual interaction could be required.

Examples:

Estimation of lens distortion

Digital camera: Olympus C1400
Zoom lens set to wide angle (9.2 mm)
Estimated lens distortion parameter: k1 = -1.63 (sigma: 0.06) x 10^-3 mm^-2
Principle point and focal length cannot be estimeted due to the orientation of the image

Original image of a building facade.

Automatically extracted image lines in green.

Image corrected for lens distortion.

Camera calibration and object reconstruction from a single photograph by Albrecht Meydenbauer: click here

Publication:

F.A. van den Heuvel, 1999.

Estimation of interior orientation parameters from constraints on line measurements in a single image

International Archives of Photogrammetry and Remote Sensing, Vol. 32, Part 5W11, pp. 81-88

(download the pdf file; 1.2Mb)