A Vehicle License Plate Recognition System Using Morphological ROI (Region of Interest) Map Generated from Morphology Operation

2017 International Conference on Control Engineering and Artificial Intelligence

计算机科学

Kim, Taewan^1 ; Kim, Eungtae^1

Dept. of Electronics Engineering, Korea Polytechnic University, 237 Sangidaehak-ro, Siheung-si, Gyeonggi-do, Korea, Republic of^1

关键词: Complex background;    Morphology operations;    Region of interest;    Standard deviation;    Threshold levels;    Vehicle license plate recognition;    Vehicle license plates;    Vertical edges;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/806/1/012004/pdf
DOI  :  10.1088/1742-6596/806/1/012004

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

In this paper, morphology features are used to improve license plate region and Morphological ROI (MROI) map that consists of a standard deviation map of open and close images is proposed. Unlike conventional morphology techniques, MROI map is used to improve the detection of license plate in a complex background. Furthermore, distribution of the MROI map was analyzed to determine a threshold level, generate binary code images, and finally, determine the vehicle license plate region by using the features of the plate such as aspect ratio and vertical edge intensity. The proposed method could improve the error rate by approximately 4-7% depending on the pattern of the images.