文档介绍：中南民族大学硕士学位论文汽车牌照定位与字符分割的研究及实现姓名:李国立申请学位级别:硕士专业:计算机应用技术指导教师:段汕 20090523 汽车牌照定位与字符分割的研究及实现 II 摘要车辆牌照自动识别是智能交通系统中的关键技术。在高速公路收费,车辆检测,停车场监控与管理,路面行驶车辆监控等领域有着广泛的应用前景。本文对车辆牌照识别系统中汽车牌照的定位以及字符分割进行了研究,具体内容包括: 图像预处理,车牌定位及字符分割。首先, 对图像预处理进行研究,分别对图像进行灰度化、灰度拉伸、二值化、边缘检测等方面的研究。对采集到的图像进行灰度化以及直方图灰度均衡化处理, 以改变在光照不理想的情况下引起的对比度过大或过小对图像的影响。然后在图像二值化方面使用了 Otsu 算法进行图像的二值化,其二值化得到较好的效果。再深入分析了在图像预处理方面常用的边缘检测算子,并选用 Sobel 算子作为本文的边缘提取算子。其次,进行牌照定位算法的研究。先介绍了车牌定位常用的方法,然后提出了一种基于数学形态学与几何形状特征相结合的车牌定位方法。该方法利用数学形态学的闭开运算将牌照字符区域分成独立的连通域,并通过牌照的几何特征: 宽度和高度比例及面积范围来实现车牌的准确定位。实验结果表明这种定位方法准确率较高,运行速度较快。最后,研究牌照字符分割。定位后的牌照预处理包括牌照倾斜角度的校正和边框及铆钉去除。在字符分割时,利用垂直投影法进行分割,在分割的过程同时提出了参照车牌字符的个数及字符的宽度和高度及字符之间的间隔来分割,并给出了这些参数取值一个可行的范围值。实验结果表明,这种字符分割的方法效果准确,尤其是字符之间存在粘连或噪声干扰的车牌,也能起到较好的分割效果。关键词:二值化;边缘检测;数学形态学;车牌定位;字符分割中南民族大学硕士学位论文 III Abstract Vehicle License Plate Recognition is a key technology of intelligent transportation System . Charges on the highway, vehi cle detection, monitoring and management of car parks, Road vehicle monitoring in areas have broad application prospects. License plate location and license plate character segmentation of the license plate recognition system ,is mainly studied in this paper. Firstly, study of image preprocessing,gray-scale images, respectively, and gray-scale stretch, binarization, edge detecti on, etc. were studied. gray and gray-scale histogram equalization processing of the collected images ,to e non-ideal in the light of the contrast caused by too large or too small to image the impact of effective information .and then use the Otsu algorithm for image binarization , it get a better binarization results. Further analysis monly used in image processing edge detection operator,then select the Sobel operator for the edge operator of this article. Secondly, license plate location to study, first introduced the monly used in license plate location, And then the license plate location method based on mathematical morphology is given. The method is to images of preprocessi