Design of real time vehicle identification system

Abstract

One of the vehicle identification techniques is to recognize the license plate numbers. The license plate recognition system consists of three major parts: sensing, image capturing, and recognition. The car sensing part is usually based on proximity or loop coil sensors and sends the signal to the image capturing part to trigger a video camera. The image capturing part, after receiving the trigger signal from the sensing part, captures the car image using high speed shutter to reduce motion blurring. The recognition part then processes the captured image and recognizes the license plate numbers. In this paper, we describe new methods for each part. For the sensing and image capturing, we propose a sensorless technique based on temporal filtering. For the recognition part, we present a robust algorithm for locating the license plate in a car image. >