For detecting and identifying license plate, color and type for vehicle and motorcycle, the system processes the pictures that are coming from road monitoring camera. The system is able to work and response in a real time without any delay, moreover accuracy exceeds 95%. The system can apply to highway and general road to supervise and seize illegal vehicles.

In conventional image techniques, although it can find license plate position, it doesn’t differentiate the type of license plate and only identify on single lane. The system applies to highway and general road, in addition, the system can identify license plate on multi-lane. The system used deep learning to identify license plate、color and type and proposes a post processing for license plate which find out each word to assemble and differentiate plate number to find the best result and save information into the database. The system is able to work and response in a real time without any delay, moreover accuracy exceeds 95%. As of now the road monitoring camera is managed by police and Ministry of Transportation and Communications. The system can apply to police department, Ministry of Transportation and Communications, apartment complex and parking lot. We can obtain time quickly by vehicle from road monitoring camera to accelerate case handling time and reduce the cost of human resources.

一般道路、高速公路、社區、戶外與室內停車場。