UGent - 2014 met Thomas Planckaert, Sam Vloeberghs, Wouter Deryckere,Benjamin Cottyn en Sebastiaan Dumoulein C++, OpenCV
Most visually impaired people are able to memorize only a few trajectories with suﬃcient detail. It would be a big step forward if the amount of trajectories could be expanded by using a vision based navigation system, which will be described in this paper. By using the opencv library, a route is split into segments from which features are extracted. By training a Support Vector Machine (SVM) with these features, this SVM will provide data that will be inserted into a Markov chain. This automate will then use all the collected data to estimate the position of the user on route.