1. Modal-Based Camera Correction for Large Pitch Stereo Imaging
- Author
-
Nathaniel J. Short, Prather Lanier, Lynn Abbott, and Kevin Kochersberger
- Subjects
business.industry ,Computer science ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,Astrophysics::Instrumentation and Methods for Astrophysics ,Stereo imaging ,Stereopsis ,Camera auto-calibration ,Computer Science::Computer Vision and Pattern Recognition ,Distortion ,Calibration ,Computer vision ,Artificial intelligence ,business ,Stereo camera ,Computer stereo vision ,Camera resectioning - Abstract
Stereo imaging is typically performed using two cameras that have been calibrated to account for lens-induced distortion and pointing errors, resulting in rectified images that are processed to obtain distance information. The accuracy of a 3-D map obtained from stereopsis is closely tied to the calibration data, and so relative motion between the cameras must be kept small. In order to reduce errors from a stereo image caused by motion, the structural connection between the cameras can be stiffened, but this comes with a weight and size penalty. For cameras that have a large baseline (pitch) distance, it may be impossible to have enough stiffness in the structure to obtain reasonable error bounds. An alternative approach is to model the camera motion using a modal technique and account for this motion during imaging. This paper outlines a procedure for stereo camera correction using measured accelerations to optimally trigger the camera. Results of the technique are shown for a simple beam that is center-mounted to a shaker to induce symmetric bending.
- Published
- 2011
- Full Text
- View/download PDF