| A correlation-based technique is proposed for the determination of camera magnification and scale-factor parameters. Experimental verification was conducted by transferring the high precision of a commercial translator, which can be 10-5 or smaller in fractional error, to the image plane of a camera system. Key to this technique is the averaging of pixel boundary effects with a large-area planar random-dot target. The correlation method was expected to be relatively insensitive to camera random errors, target characteristics, illumination non-uniformities and localized imperfections in the camera or the target. Variations of the focal length across the camera field-of-view were measured as well. Systematic errors, due to lens aberrations, target and translator tilt and rotation, limit the accuracy of the method. A guide to parameter selection in light of systematic errors is presented. The technique is adjustable to the desired degree of precision, can be automated and is applicable to any imaging system built from off-the-shelf components. Future directions for increased precision and robustness are suggested. |
