The Study of License Plate Numbers Recognition Using Plastic Perceptron Neural Network

• Main Authors: Yih-Bin Yu, 余益濱
• Other Authors: I-Chang Jou
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/42576095166761160028

碩士 === 國立高雄第一科技大學 === 電腦與通訊工程所 === 91 === In this thesis, we propose the techniques of the combination of the digital image processing and the Plastic Perceptron Neural Network (PPNN) for the implementation of the License Plate Numbers Recognition System. The framework of the PPNN in this thesis is based on the Back-Propagation Neural Network (BPNN), it improves the problems of the BPNN, such as time-consuming learning speed, difficult convergence, and overall retraining when deleting patterns or adding new ones. This thesis involves some image processing techniques, such as thresholding、noise elimination、image segmentation and normalization. In order to have a high recognition rate, the characteristic vectors combine the White Run-Length and Pixel Density. The final test shows license plates recognition rate of 89% and characters recognition rate of 98.17% are accomplished.