Multi-Sensor Classification using a boosted Cascade Detector

This paper provides a description of a new low-level feature-fusion approach for real-time object recognition utilising an arbitrary number of imaging sensors and based on a boosted cascade of simple features. The approach is demonstrated by means of a vehicle detection system. The application utilises laser scanner responses for hypotheses generation and low-level features from both colour and far infrared images for hypotheses classification. A first evaluation shows promising results.