Object discrimination and tracking in the surroundings of a vehicle by a combined laser scanner stereo system

Kurzfassung

The use of sensor data for observing the surrounding environment of a vehicle is becoming increasingly popular. Especially for detecting dangerous situations, which occur too fast for the human senses, sensor systems are needed. In the following paper such a sensor system consisting of a stereo camera and a multilayer laser scanner mounted in front of a test vehicle is introduced. Both sensors are used to detect and track obstacles and other traffic objects independent from each other for future data fusion. An overview of the complete process for the object discrimination including a novel approach for a sensor cross calibration and a new method for the object refinement and the object tracking is given. The effectiveness of the algorithms are tested with real road reference data, obtained through highly precise GPS data.