Precise Vehicle Ego-Localization Using Local Feature Matching of Pavement Images

Precise vehicle localization is critical for various ITS applications. The Global Positioning System (GPS) and the Strap-down Inertial Navigation System (SINS) are two common techniques in the field of vehicle localization, but accuracy, reliability and real-time performance of these two techniques can not satisfy the requirement of some critical ITS applications such as collision avoiding, vision enhancement and automatic parking. Therefore a precise vehicle ego-localization method based on image matching was proposed, which included 3 steps: (1) Extraction of Feature Points; (2) Eliminate mismatch points; and (3) Matching of Feature Points and Trajectory generation. Through matching and validating the extracted local feature points, the relative translation and rotation offsets between two consecutive pavement images were calculated and vehicle trajectory was generated. Results show that the studied algorithm has an accuracy at decimeter-level, fully meeting the demand of the lane-level positioning in some critical ITS applications.