Effects of License Plate Attributes on Automatic License Plate Recognition

This paper describes an experiment designed to develop a comprehensive and thorough understanding of the readability of North Carolina's license plates with an automatic license plate recognition system. This research focused on law enforcement applications and used two infrared camera systems for data collection in a controlled environment involving more than 900 license plates. The field test was conducted in a controlled environment under the following settings and conditions: a test vehicle traveling at 25 mph, nighttime evaluation, closed test track, 40 ft of spacing of license plates, 9 ft of lateral offset between the license plates and camera, standard issue and specialty plate types, standard syntax and personalized plates, and various license plate ages and conditions. The key finding of this research project is that the current, standard issue license plate with characters in blue ink has the highest capture and read rates among the plates tested in this study. Factors that decreased the capture and read rates were personalized syntax, specialty license plates, and the presence of stacked characters on a specialty license plate.