Ego-Lane Position Identification With Event Warning Applications

This paper proposes a high-performance advanced driver assistance system that analyses driving scenes based on monocular cameras. The system identifies the ego-lane position and indicates if the vehicle is driving on an inner or outer lane. Dense optical flow analysis is performed and a fuzzy system is designed to achieve ego-lane position identification. An event warning application is implemented based on a hierarchical classifier and the results of ego-lane position identification. The proposed event warning system accurately issues events without having to detect vehicles first, making the system more responsive to potential approaching dangers. Also, the proposed system has a comprehensive ability to generate warnings on various types of events. The experimental results validate the effectiveness of the proposed schemes.