The preceding chapters covered most of the details relevant to our qualitative motion understanding approach, in particular the FuzzyFocus of Expansionand theQualitative Scene Model.In the following, we show the operation of this concept with two major examples using image sequences that are typical for autonomous land vehicle navigation. The first experiment was performed with an animated image sequence in order to have the ground truth and the vehicle motion parameters available. This setup was used to develop and debug the reasoning mechanism. The second example shows the results from a real image sequence taken from the moving ALV.
Unable to display preview. Download preview PDF.