A Study for Reconstruction Low-Resolution License Plate Image Base on Improving Image Alignment

• Main Authors: Chih-Ting Chen, 陳治廷
• Other Authors: Shin-Chieh Lin
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/42336201079576595159

碩士 === 國立清華大學 === 動力機械工程學系 === 94 === In this paper, we use Super-resolution to save the problem of low resolution image in license plate recognition system. Image registration is an important part of image enhancement. In the same algorithm of the image enhancement, a precise sub-pixel image registration has a better result of image reconstruction. So, we can understand the important of the image registration. There are two kinds of methods for image registration. One of the methods is feature based, the other is area based. The method of feature based finds control point first and finds the orientation of image by using control point. The method of area based finds the orientation of image by almost all over the range of image. We will design an image which's minimum shifted pixel is 0.1 pixel and minimum rotated angle is 0.1 degree. In this paper, we will find the best method of image registration in the biblliography by using the images which we designed. We also propose a new method to improve the method of feature based. We will recognition the low resolution license plate image by using the best method of image registration and try to find the information of the license plate image.