Super-resolution of license-plates using frames of low-resolution video

Abstract

In this paper, we consider the problem of multi-frame super-resolution for license plate images given low-resolution frames. Optical Character Recognition (OCR) systems may not work well when the resolution of a license plate image is low. Using high-resolution cameras is an expensive solution to this problem. However, super-resolution methods can reconstruct a high-resolution image from a set of low-resolution frames. In this paper, we propose a new interpolation method that enhances the resolution of the license plate, using consecutive frames of a video. The images obtained by this method are suitable for a license plate recognition system. Experimental results show that applying the proposed method on low-resolution frames, improves the quality of the license plate significantly, and produces better results when fed into an OCR system.