导读： 中文标题 汽车牌照识别技术的研究 英文标题 Research On License Plate Recognition Technology 作者 马玺 专业与研究方向 测试计量技术及仪器 智能化仪器及计算机测控 指导教师 胡泽 申请学位等级 硕士 机构 西南石油大学 论文完成年度 2008 中图分类号 TP3...
中文标题
汽车牌照识别技术的研究
英文标题
Research On License Plate Recognition Technology
作者
马玺
专业与研究方向
测试计量技术及仪器 智能化仪器及计算机测控
指导教师
胡泽
申请学位等级
硕士
机构
西南石油大学
论文完成年度
2008
中图分类号
TP391.41
中文关键词
车牌识别；车牌定位；字符分割；车牌字符识别；神经网络；Sobel算子
英文关键词
License plate recognition; License plate positioning; License plate character recognition; Neural network; Sobel operator
中文摘要
汽车牌照自动识别系统是以汽车牌照为特定目标的专用计算机视觉系统，是计算机视觉和模式识别技术在智能交通领域应用的重要研究课题之一，是实现交通管理智能化的重要环节。汽车牌照自动识别系统的核心技术主要包括车辆定位、字符分割和车牌字符识别。本文着重对车牌定位，字符分割及字符识别的各种算法进行了深入的分析和比较，并提出了自己的观点。在车牌定位部分，提出了基于Sobel算子的投影定位算法，能够快速准确地定位出车牌。在字符分割部分，着重对图像二值化技术，图像倾斜矫正技术，字符分割技术及字符归一化技术进行了研究，并利用垂直投影的方法有效的对车牌字符进行了分割。在字符识别部分，提出了基于BP神经网络的识别算法用于车牌字符识别，并得以实现。
最后基于上述算法编制出车牌识别软件，使用实际车牌图片对软件进行了测试。测试结果表明本文设计的车牌处理算法一方面提高了车牌处理的正确性，另一方面在一定程度上减少了车牌处理时间，因此更加适合应用于实际的车牌识别系统中。
英文摘要
License Plate Auto Recognition System is special computer vision system which particular objective is license plate. It is one of the important research topic in intelligent transportation field. It’s the significant part to achieve transportation management intelligentization.The kernel technique of the License Plate Auto Recognition System is mainly composed of license plate locating, character segmentation and character recognition. This paper analyzes and compares various algorisms of plate locating, character segmentation and character recognition. In the part of license plate positioning, the author presents projection locating algorism which based on Sobel operator. This can quickly and accurately positions the license plate. In the part of character segmentation, the author mainly researches image binary processing technology, incline image correction technology, character segmentation technology and character normalization technology. And the author uses vertical projection method to segment license characters effectively. In the part of character recognition, the author proposes recognition algorism which based on BP neural network, and it achieved successful.
At last, the author develops a License Plate Recognition Software based on these algorisms, and uses various actual license plate images to testing. Test results show that this plate processing algorisms in this paper increased the accuracy of plate processing on the one hand. On the other hand, it reduced the license plate processing time to a certain extent. Therefore, it more appropriate to apply to the actual license plate recognition system.
目录/全文下载
[PDF]
