A Study on Intelligent Vehicle Plate and Model Recognition

• Main Authors: LIN, SEN-MAO, 林森茂
• Other Authors: Chou, Wen-Kuang
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/33541926265611018001

碩士 === 靜宜大學 === 資訊碩士在職專班 === 104 === The vehicle license plate and model recognition is a very popular research topic in decades. The system is mostly applied in the parking management, and also in the police for investigation of the stolen vehicles and road surveillance. However, the application on the police is still not reliable for variable and diverse backgrounds. Compared with the parking management, the police applications are much more complicated. This study mainly proposes and implements the method based on grayscale image to enhance the accuracy and the speed of recognition even under the noisy environments. The experiments use 800 pictures to test the correctness of license plate localization, plate character recognition, and vehicle make comparison and recognition. The experiment results show that 95% correct rate for license plate recognition under complicated backgrounds excluding bad weather conditions and 100% correct rate for specific vehicle model comparison if the targeted objects and compared samples have similar views.