基于数学二值形态学的车牌定位与字符分割

License Plate Localization and Character Segmentation Based on Mathematics Binary Morphology

本方法是基于数学二值形态学的车牌定位和字符分割的方法。在车牌定位中,首先提取图像的边界,腐蚀非车牌边界,用矩形结构元素连通剩下的区域,去除最大面积以外的其它区域。最大的连通区域即为车牌所在的位置。在字符分割中,首先用双线性插值将车牌图像调整到固定大小,确定字符的上下边界,然后逐个判断字符的具体宽度,用于分割字符。通过matlab 7.0的验证,该方法解决了车牌褪色,车牌开裂,非连通字符和连续的数字1等对自动识别系统的影响问题。

this way is about the way of the License plate localization and character segmentation based on Mathemat ics binary morphology. There are four steps in the license plate localization. Firstly, get the edge of the image; sec ond,erode some edges that can＇ t be the edge of the license plate;thirdly, handle the others with the rectangular structure element ;fourthly ,delete the areas except the biggest area. The biggest area is the license plate. There are three steps in the character segmentation. Firstly, adjust the vehicle license plate image to the fixed size with the bi linear interpolation; secondly, get the lower and upper bounds of the characters, thirdly, get the width of the charac ter one by one to segment the character. This way can be used to handle the questions about the fade of the license plate, the crack of the license plate, non communicating characters and the consecutive number 1. The tool is matlab 7.0.