摘要
为减少计算机对车牌图像识别的复杂性,在图像识别前,要对车牌图像进行倾斜度的调整。针对二值化后的车牌图像进行特征分析,利用垂直投影,计算出字符间的距离,根据倾斜车牌和非倾斜车牌垂直投影的字符间距离大小的变化,确定图像的倾斜角度;利用图像旋转的算法,得到非倾斜的图像,从而完成车牌的倾斜度的调整。基于效率的考虑,又提出了该倾斜算法的优化算法。
In order to decrease the complexity of identifying the image of vehicle license with computer, we should adjust the gradient of image before identifying. The paper analyzed the character of two- valued vehicle license by vertical projection, and computed the distance between two characters. According to the change of distance which comes from the vertical projection towards lean image and horizontal image, we can find out the lean angle of image. By the rotatation, we can get a horizontal image. Considering efficiency, we also preferred out an optimum algorithm.
出处
《济南大学学报（自然科学版）》
CAS
2007年第3期246-248,共3页
Journal of University of Jinan(Science and Technology)
基金
济南大学科技发展基金(Y0613)
关键词
数字图像
倾斜调整
车牌识别
numeric image
adjustment of gradient
identification of license
