A-PNR

Automatic Plate Number Recognition (APNR) has important applications in traffic surveillance, toll booth, protected parking lot, No parking zone, etc. It is a challenging problem, especially when the license plates have varying sizes, the number of lines, fonts, background diversity etc. This work aims to address APNR using deep learning method for real-time traffic images. We first extract license plate candidates from each frame using edge information and geometrical properties, ensuring high recall using one class SVM. The verified candidates are used for NP recognition purpose along with a spatial transformer network (STN) for character recognition. Our system is evaluated on several traffic images with vehicles having different license plate formats in terms of tilt, distances, colors, illumination, character size, thickness etc. Also, the background was very challenging. Results demonstrate robustness to such variations and impressive performance in both the localization and recognition.