Object Oriented Motion Estimation in Color Image Sequences

Abstract

This paper describes a color region-based approach to motion estimation in color image sequences. The system is intended for robotic and vehicle guidance applications where the task is to detect and track moving objects in the scene. It belongs to the class of feature-based matching techniques and uses color regions, resulting from a prior color segmentation, as the matching primitives. In contrast to other region-based approaches it takes into account the unavoidable variations in the segmentation by the extension of the matching model to multi matches. In order to provide extended trajectories, color regions that could not be matched on the feature level are matched on the pixel level by the integration of a correlation-based mechanism. The usage of color information and the combination of feature-based and correlation-based matching leads to robust and efficient algorithms. The system was applied to a motion segmentation task in vehicle guidance. Experiments on more than 1000 natural color outdoor images, taken from a moving car, show promising results.