Options
1989
Conference Paper
Titel
Interpretation of traffic scenes by evaluation of optical flow fields from image sequences
Abstract
TV cameras are very useful to record traffic scenes due to their high resolution compared to their size and costs. In comparison with single images, image sequences comprise information about the dynamic aspects of the recorded scene. Dynamic aspects play a crucial role in traffic scenarios. Traffic scenes with different complexity can be characterized by the structure of the scene. Depending on the structure of the scene different tasks could be solved by image sequence evaluation. In this contribution two approaches for the evaluation of optical flow fields from image sequences are discussed. The first approach describes the detection and tracking of moving objects on a complex street intersection. The second approach presents an algorithm for the detection of obstacles on the road of an autonomous vehicle. It will be shown how two-dimensional optical flow fields can be interpreted to infer information about the three-dimensional environment. Results obtained from investigations with real world image sequences will be presented.