License plate localization using a statistical analysis of Discrete Fourier Transform signal

• Statistics of Discrete Fourier Transform.
• Five features are chosen to represent plate signal.
• Bayesian approach is used to classify the plate region from other regions.
• The method achieves classification rate of 97.27%.
• Comparison analysis is provided.

License Plate Recognition (LPR) is a well-known problem and it has developed as a coherent framework. Research continues on the topic due to the diversity of license plates and outdoor illumination conditions which require attention. One of the most important steps in LPR is the localization part where license plates are extracted from video captured images. In this article we introduce a new approach of plate localization using a statistical analysis of Discrete Fourier Transform of the plate signal. The plate signal is represented by five statistics: strength of the signal, normalized maximum amplitude, frequency of maximum amplitude, frequency center and frequency spread. Combining with the color-based histogram thresholding, the method achieves 97.27% accuracy using plate signals from binary images. Comparative analysis is also reported.