摘要
车牌字符分割与识别是智能交通系统中的重要环节,而倾斜变形的车牌图像对车牌字符分割与识别有很大影响,为解决这个问题,提出了一种基于方向场的车牌校正方法。建议方法采用Sobel算子计算车牌图像的梯度方向场,通过对方向场角度直方图的分析找出车牌水平倾斜角,根据水平倾斜角对车牌图像进行水平旋转校正并二值化。然后,应用垂直投影法对车牌进行竖直方向上的变形校正。建议方法被用来对200幅各种情况的车牌图像进行校正,效果较好。
The character segmentation and recognition are important steps in the intelligent transportation sys- tems. The skew distortions of license plate have a great influence on the accurate character segmentation and recog- nition. To solve the problem, a novel approach based on orientation field for skew correction of license plate is pro- posed. First, gradient orientation field of a plate image is estimated by Sobel operator, and it converts orientation angles and counts its histogram in order to get the horizontal incline angle. The plate image is rotated with this angle and binarized. Then, the vertical distortion of license plate image is corrected by the vertical projection method. The proposed method is tested with 200 plate images. The experimental results show that the proposed method is ef- fective.
出处
《贵州科学》
2011年第2期85-88,共4页
Guizhou Science
基金
贵州省高层次人才特助经费(黔人领函[2009]03)
贵州民族学院自然科学类学生重点资助科研项目(2008XS28)
关键词
车牌倾斜校正
方向场
方向角直方图
垂直投影
correction of license plate, orientation field, orientation angle histogram, vertical projection
