License plate detection for multi-national vehicles – a generalized approach
- 39 Downloads
License plate detection for vehicle identification is one of the key problems for traffic surveillance in urban areas. Most of the existing methods that can handle multiple license plates are country-specific as color information has been typically targeted. In this paper, we propose a real-time multiple license plate detection method feasible for multi-national vehicles having variable colors, sizes and geometrical attributes. A region-of-interest is initially identified for each vehicle using a fuzzy inference system based on the salient feature of its rear lights as license plates generally exist in a vicinity of these lights. Due to the abundance of edges within the license plate region, a local recursive analysis approach is utilized to locate the license plate candidate within each region-of-interest after tilt correction using a rear-light alignment technique. To verify the detected region as a true license plate, a unique combination of local image features is used to achieve high precision. The proposed method has been tested on 2200 images taken during various weather and illumination conditions to detect 5379 license plates out of 5945 available vehicles with 90.5% accuracy. The proposed approach outperforms the conventional and deep learning methods to achieve superior performance with the ability of being applied to multi-national vehicles.
KeywordsIntelligent vision system License plate detection Local image features Multi-national vehicles Vehicle identification Vehicle rear lights
The authors would like to thank Dr. Bin Tian from Institute of Automation, Chinese Academy of Sciences, Beijing, China for providing the dataset for experimentation.
Conceptualization, Muhammad Rizwan Asif; methodology, Muhammad Rizwan Asif; software, Muhammad Rizwan Asif and Tiexiang Wang; validation, Muhammad Sadiq Fareed and Subhan Khan; formal analysis, Muhammad Sadiq Fareed and Subhan Khan; resources, Chun Qi; data writing—original draft preparation, Muhammad Rizwan Asif; writing—review and editing, Muhammad Rizwan Asif and Chun Qi; supervision, Chun Qi; funding acquisition, Chun Qi.
This research is supported by the National Natural Science Foundation of China (Grant No. 61572395 and 61675161).
Compliance with ethical standards
Conflict of interest
The authors declare no conflict of interest.
- 11.Le, TS, Tran VN, Hamamoto K (2014) A robust and flexible license plate detection method. Proc IEEE Intl Conf Advanced Technologies Communication, p 326-331Google Scholar
- 12.Li H, Yang R, Chen X (2017) License plate detection using convolutional neural network. IEEE Intl Conf Comp Comm:1736–1740Google Scholar
- 13.Li J, Liang X, Shen S, Xu T, Feng J, Yan S (2018) Scale-aware fast R-CNN for pedestrian detection. IEEE Trans Multimedia 20(4):985–996Google Scholar
- 14.Li-Hung W, Chao-Kai C, Chung-Bin W (2014) A real-time architecture of multiple features extraction for vehicle verification. Proc IEEE Asia Pacific Conf Circuits Systems, p 615–618Google Scholar
- 16.Peng J, Wang N, El-Latif AAA, Li Q, Niu X Finger-vein verification using Gabor filter and SIFT feature matching. Eighth Intl Conf Intell Inf Hiding and Multimed Sig Proc, p 45-48.Google Scholar
- 17.Prabhakar P, Anupama P (2014) A novel design for vehicle license plate detection and recognition. Proc IEEE 2nd Intl Conf Current Trends in Engineering and Technology, p 7–12.Google Scholar
- 26.Zeiler MD, Fergus R (2014) Visualizing and understanding convolutional networks. Proc Euro Conf Comp Vis, p 818–833.Google Scholar