Much research on automatic license plate recognition system has focused on applications such as traffic violation detection, stolen vehicle identification, etc. These applications are implemented on a stationary platform where the environmental parameters (i.e. light conditions, location and orientation of the object) of the deployment site are obtainable. Stationary-based implementations are tuned to perform best under a predetermined environment. In mobile environments, the accuracy of these systems drops significantly. Here, the issues related to environmental parameters for mobile environments are presented. Components studied as a part of a mobile license plate recognition system include special image acquisition procedures, hardware-based image processing and character image recognition techniques. A new mobile system is proposed and tested. Finally, the real-time implementation requirements are presented based on a performance and complexity analysis for such a system. |