MoBT5.3 - Fast Super-resolution for License Plate Image...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Fast Super-resolution for License Plate Image Reconstruction YUAN Jie 1 , DU Si-dan 2 , ZHU Xiang 3 Department of Electronic Science and Engineering, Nanjing University, Nanjing 210093, China 1 yuanjie@nju.edu.cn, 2 dsd@ese.nju.edu.cn, 3 zhuxiang0703@gmail.com Abstract A fast super-resolution reconstruction algorithm designed for license plate recognition is proposed in this paper. It uses a new reduced cost function to produce images of higher resolution from low resolution frame sequences. Computational cost required in this algorithm is much lower compared with other methods. The effectiveness of the proposed algorithm is demonstrated through blind reconstruction experiments with real videos, whose result images are nearly equivalent to those yielded by classical MAP-based approaches. The presented algorithm can be applied in real-time recognition systems to improve their performances, and to reduce the requirement of imaging hardware. 1. Introduction Super-resolution reconstruction (SR) aiming to form a high resolution image by combining multiple low-resolution images has received much attention these years. By now one main obstacle to the real application of this technology is the lack of accurate motion estimation of complex movements in general videos. However, under certain circumstances a simple global motion model could meet our needs. One example is the License Plate Recognition (LPR) field, in which the movement pattern of object is comparatively simple for estimating. Most LPR algorithms and commercial LPR systems are designed for toll gates, parking lots, or other situations, where it is easy to take sharp and large plate photographs from vehicles that stay still or move in a very low speed. Unfortunately, if we employ such systems to identify faster-moving vehicles, plates might hardly be recognized because we need to consider motion blur problems then. For running vehicles, the shorter the distance between the camera and the object is, the more severe the motion blur becomes. On the other hand, if cameras were stationed far away or were zoomed out to avoid blur effect, the reduced resolution of plate images would make it even harder to recognize. Another problem concerns additional information attached to the plate, such as province initials (as in China, these are Chinese characters), country flag (as in Europe), or state identification (as in USA). The recognition of such information would require even higher quality of images captured for LPR. Some institutions, including leading companies in this field, developed high-speed systems for fast moving vehicles. However, these systems depend largely on sophisticated imaging hardware with much higher resolution and shutter speed to tackle the problem above. They are, of course, very expensive, and have hindered the extensive use of LPR system for automatic traffic monitoring.
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 4

MoBT5.3 - Fast Super-resolution for License Plate Image...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online