András Bódis-Szomorú

Specular surface reconstruction for multi-camera corneal topographer arrangements

A novel specular surface reconstruction approach developed for and used in an experimental multi-camera corneal topographer arrangement is presented in the paper. The tear-film coated, and therefore specular, corneal surface is mathematically reconstructed from the corneal images taken simultaneously by the calibrated cameras of the arrangement. The surface reconstruction is achieved by the joint solution of partial differential equations (PDEs) that is derived from the mapping between the measurement-pattern and its virtual images taken by the cameras. The PDEs - each given in an advantageous form for one of the cameras - describe the phenomenon of light-reflection at the corneal surface. The portions of the corneal surface that reflect - possibly different - measurement-pattern points to more than one camera convey starting points for the mathematical reconstruction of the surface.

A Far-Range Off-line Camera Calibration Method for Stereo Lane Detection Systems

A far-range camera calibration method for vision sensors used in automatic lane detection systems is investigated. Autonomous vehicle guidance based on stereo image acquisition requires the knowledge of the camera parameters, such as camera position, orientation, lens distortion, focal length etc. with high precision. Calibration is performed by placing markers in known positions in front of the vehicle holding the camera(s) and by detecting the location of their images in the acquired snapshots. The parameters are then estimated by minimizing the geometric error in the image. Most of the methods discovered in the literature do not consider errors in the measurement of marker locations. However, a precise marker arrangement for far-range calibration might be expensive to set up. This paper shows that imprecision in the 3D location of the markers can not be tolerated in some practical situations and in such cases, a different optimization is proposed in the calibration of the extrinsic camera parameters (camera position and orientation). A position estimation method is also shown to extract marker positions from far-range laser distance meter measurements. The estimated marker positions are corrected by considering aiming errors, as well. Parameter errors are also estimated. It will be shown that our new calibration and optimization method is expected to result a good accuracy compared to commercial methods.

Optimization methods to calibrate a stereo rig with increased accuracy for vehicular applications

In this paper, several methods are presented to fully calibrate a stereo vision system for far-range outdoor applications. Traditionally, intrinsic camera parameters are determined from a high number of feature points, while pose estimation is solved with a few (tens). Close-range setups are not well suited for far-range applications, while far-range setups can easily deviate from planar and small measurement errors may result in large errors in the marker positions. Our Maximum Likelihood (ML) formulation to the stereo pose problem takes such inaccuracies into consideration. Herein, our earlier results are extended by decoupling pose estimation into an inter-camera and a rig-to-world pose problem to avoid relying on a small number of features with all extrinsic parameters. A high number of point-matches are extracted from image pairs acquired during on-line operation, and are incorporated into the procedure for off-line inter-camera pose estimation via the essential matrix. Using real and synthetic data, it is shown how the far-range arrangement and the matches can be used to adjust the camera models, even after a sound intrinsic calibration and ML pose estimation. The proposed methods rely on point matches between views and, thus, are expected to produce better reference for the evaluation of autocalibration methods that typically begin with establishing such correspondences.