Hinderniserkennung für mobile Arbeitsmaschinen durch Fusion von Radar- und Videodaten

Abstract

This article introduces an automatic obstacle detection method based on radar and mono-camera fusion. The obstacle detection is carried out by segmenting the traversable ground plane in the camera image. The segmentation is based on the so-called watershed transformation. The coordinates of obstacles detected by the radar sensor, are transformed to image coordinates and used by the watershed transformation as background markers.