Robust Orientation, Calibration, and Disparity Estimation of Image Triplets

Abstract. This paper addresses robust automatic orientation, calibration, and disparity estimation for generating visualizations from image triplets. Here, robust means, that meaningful results are obtained for a larger number of triplets without changing any parameter. This is achieved, e.g., by using as initial search space the whole image and by automatically estimating the search width for disparity estimation. The approach works for wider baselines than standard approaches for sequences. Results for visualization based on the trifocal tensor show the validity of the approach.

• Type: Conference
• Year: 2003
• Where: DAGM
• Authors: Helmut Mayer
• Keywords: DAGM 2003, Disparity Estimation, Image Triplets, Robust Automatic Orientation