Research Article
An Incomplete License Plate Image Intelligent Recognition System Based on the Generated Counter Network
@INPROCEEDINGS{10.1007/978-3-030-94185-7_38, author={Mi Meng and Chun-hu He and Xiao-jing Qi}, title={An Incomplete License Plate Image Intelligent Recognition System Based on the Generated Counter Network}, proceedings={IoT and Big Data Technologies for Health Care. Second EAI International Conference, IoTCare 2021, Virtual Event, October 18-19, 2021, Proceedings, Part I}, proceedings_a={IOTCARE}, year={2022}, month={6}, keywords={Generative counter measure network Incomplete license plate image Intelligent recognition}, doi={10.1007/978-3-030-94185-7_38} }- Mi Meng
Chun-hu He
Xiao-jing Qi
Year: 2022
An Incomplete License Plate Image Intelligent Recognition System Based on the Generated Counter Network
IOTCARE
Springer
DOI: 10.1007/978-3-030-94185-7_38
Abstract
In order to solve the problems of high recognition accuracy and long time-consuming recognition in the traditional incomplete license plate image recognition system, the paper proposes an incomplete license plate image intelligent recognition system based on a generative confrontation network. According to the composition of the recognition system, the framework of the intelligent recognition system for incomplete license plate images is designed. The hardware platform of the system is the S3C6410 embedded development board, which is based on Samsung’s ARM11 processor chip, and the CPU is based on the ARM111176JZF-S core design. A powerful multimedia processing unit is integrated inside, which supports hardware encoding and decoding of video files in formats such as Mpeg4, H.264/H.263, and can be output to LCD and TV display at the same time. The 6410 uses a high-density 6-layer board design, which integrates 256M DDR RAM, 256M/1 GB SLC Nand Flash memory. The system software uses a generative confrontation network to repair the incomplete license plate image, and combines the template matching method to complete the license plate image recognition. Experimental results show that the designed intelligent recognition system has high recognition accuracy and can effectively reduce the time-consuming recognition.
