High-Efficient License Plate Recognition System Based on Deep Learning

• Main Authors: LAI, JIAN-LIN, 賴建霖
• Other Authors: KUO, PO-YU
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/u9nxw5

碩士 === 國立雲林科技大學 === 電子工程系 === 107 === The license plate recognition system is widely used in a modern society where urbanization is increasingly developed, such as traffic enforcement, intersection monitoring, and automatic parking lot charging. With the rapid development of artificial intelligence in recent years, neural networks have been widely used in the field of image recognition. Compared to the existing license plate recognition system, it is necessary to adjust the algorithm for different road conditions. The neural network applied deep learning to perform image recognition and combines big data to collect different road condition data to train an identification model. The identification model has the advantage of expanding labor cost and achieving similar effects. This thesis proposes a high-efficient license plate recognition system based on deep learning, in order to achieve the low complexity of the design of the license plate recognition system, while recognizing the license plate with a tilt of -60 degrees to 60 degrees and a high recognition rate. Through the image recognition, the license plate character position of the input image is detected and extracted, and the license plate character extracted and recognized by the neural network is recognized. The results show that the license plate recognition system achieves a high recognition rate through AOLP database detection and verifies the effectiveness of the system through simulation images.