#1楼主：车牌自动识别(原创 论文整理中)摘要
文章发表于：2008-03-17 01:54
中文摘要
本文通过对大量资料的搜集、整理近年来国内外在车牌定位分割领域的最新研究成果和最新进展，对车牌区域的固有特征和目前的车牌定位分割技术进行了系统的研究和探讨。本文的算法以车牌区域的纹理分析为基础，有效提高了定位的准确性。数学形态学是图像处理领域的一项新兴技术，在车牌的精确定位分割中利用形态学对二值图像做一系列的开闭等变换并根据牌照的先验知识实现了牌照的精确分割。实验结果表明该算法能够比较快速、准确、定位分割出车辆牌照。本文还对车牌图像的字符分割进行了研究，采用基于投影的分割方法取得了较好的分割结果，为字符识别打下了基础。本论文详细阐述了基于车牌纹理特征的车辆牌照定位分割算法，并在文章的最后对后续工作的前景进行了展望。
关键词:图像分割;车牌定位;数学形态学;字符分割
ABSTRACT
On the basis of the work that summarized the latest research achivements and development of license plate location and segmentation and discussed the intrinsic characteristic of license plate, this article put forward the algorithm based on texture analysis, which make the locating-of-license exactly more. In order to preciously locate the licenses, the algorithm make use of series of morphological transformation, which is a new technology of image processing, and also employ other prophetic knowledge such as the ratio of width and length, etc. The experimental results show that this algorithm is efficient in locating the licenses while not sensitive to unconstrained illumination conditions and irregular background. Besides, this article play research in character segmentation and obtain some preferable results that provide the foundation for character recognition by means of projection-based character segmentation algorithm. This paper presents the method in details and introduce the prospect of the following work at the end of the article.
Keywords: Image segmentation, License location, Mathematical morphology, Character segmentation.