Reestablishing Consistency of Uncertain Geometric Relations in Digital Images
In uncertain geometry in the 2D plane, points are replaced by uncertainty regions. By allowing uncertainty several geometric notions such as parallelism and concurrency become inconsistent with Euclidean geometry. In previous work we explained how consistency can be partially restored by graph-theoretical grouping algorithms. In this paper we study inconsistencies at a higher-level, e.g., the violation of Desargues’s Theorem or Pappus’s Theorem. We provide a simple algorithm that completely restores Euclidean consistency. Although the algorithm may not give optimal results with respect to grouping, it shows a way to develop more sophisticated algorithms to obtain global consistency in uncertain geometry.
KeywordsParameter Point Uncertainty Region Parallel Pair Digital Plane Parallel Incidence
Unable to display preview. Download preview PDF.
- P. Veelaert. Algorithms that measure parallelism and concurrency of lines in digital images. In Proceedings of SPIE’s Conference on Vision Geometry VIII, pages 69–79, Denver, 1999. SPIE.Google Scholar
- P. Veelaert. Line grouping based on uncertainty modeling of parallelism and collinearity. In Proceedings of SPIE’s Conference on Vision Geometry IX, pages 36–45, San Diego, 2000. SPIE. 268Google Scholar
- P. Veelaert. Parallel line grouping based on interval graphs, volume 1953 of Lecture Notes in Computer Science, pages 530–541. Springer, Uppsala, Sweden, 2000.Google Scholar
- P. Veelaert. Collinearity and weak collinearity in the digital plane, volume Digital and Image Geometry of Lecture Notes in Computer Science, pages 439–453. Springer, Dagstuhl, Germany, 2001.Google Scholar
- P. Veelaert. Concurrency of line segments in uncertain geometry, volume 2301 of Lecture Notes in Computer Science, pages 289–300. Springer, Bordeaux, France, 2002.Google Scholar