Image Processing for Driver Assistance

摘要: Systems for automated image analysis are useful a variety of tasks and their importance is still growing due to technological advances an increase social acceptance. Especially in the field driver assistance systems progress science has reached level high performance. Fully or partly autonomously guided vehicles, particularly road-based traffic, pose demands on development reliable algorithms conditions imposed by natural environments. At Institut fur Neuroinformatik methods analyzing driving relevant scenes computer vision developed cooperation with several partners from automobile industry. We introduce system which extracts important information taken CCD camera installed at rear view mirror car. The approach consists sequential parallel sensor processing. Three main namely initial segmentation (object detection), object tracking classification realized integration branch fusion branch. gain this given integrative coupling different providing redundant information.