Radar-vision fusion for vehicle detection by means of improved haar-like feature and AdaBoost approach

This work describes a vehicle detection system that uses fusion of vision and radar data. The radar provides a first estimation of the lateral position of vehicle candidates and the related distance information. This information is used to define a region of interest (ROI) that is subject to verification. A video camera is used for the verification purpose. The projection of the ROI onto the image plane is scanned via an AdaBoost object detection algorithm, and thus radar detection can be verified and more specific data of the vehicle´s 3D position and width can be given. Moreover, the distance information provided by radar is used to choose optimal parameters during the visual detection process, e.g. properties of the scan window and parameters for fusing detections. In addition, mutual information for haar-like feature selection is used to increase detection rates.