Vehicle anti-collision warning system is a key research area of vehicle safety. It is also a necessary means to enhance driving safety and reduce traffic accidents. The visual location method is presented for this system based on license plate image. This method realizes visual positioning with the help of the mounting points of license plate. At the same time, the feasible robust analysis method is proposed for the visual positioning method with license plate frame image. Finally, the robust analysis determines good angles for commutating rotation parameter in the method of license plate image visual positioning. At the same time, it determines good angles for translation parameter and for rotation and translation transform. Therefore, robust analysis determines positions where the visual positioning method has high positioning accuracy. Robust analysis for license plate vision positioning method is useful in analysis of high positioning accuracy positions for the proposed license plate image visual location method combined with plate mounting points. The research content of this paper is beneficial to driver decision-making and improves the safety of collision warning system and intelligent vehicle.

This work was published on International Journal of Pattern Recognition and Artificial Intelligence,2017,31(1):1-15. titled Robustness Analysis of License Plate Visual Positioning Method in Vehicle Collision Warning System.