基于分块聚类的车牌图像二值化方法  

Binary Method of License Plate Image Based on Blocking-Clustering

作　　者：唐泉[1] 樊晓平[1] 黎燕[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410075

出　　处：《计算机仿真》2010年第3期286-289,352,共5页Computer Simulation

摘　　要：利用车牌图像中车牌部分所占比例小、车牌字符与底色灰度对比度大的特点,为运算简单,易识别定位,提出了一种基于分块聚类的车牌图像的二值化方法。方法先将整幅灰度车牌图像等分成多个小块,对每个小块中的像素点根据其灰度值进行聚类,再将聚类结果二值化而得到各小块的二值化结果,合成各小块的二值化结果得到整幅车牌图像的二值图像。实验结果表明,与OTSU二值化方法相比,不仅经该方法处理得到的二值图像能很好地体现车牌特征、有效地降低车牌图像背景的复杂性,而且方法更适用于处理非正常光照条件下的车牌图像。在此基础上能有效地提高车牌定位系统的实时性和适应性。Making use of the features of license plate taking a small part of the hole license plate image and the stronger contrast of the characters with the background, this paper proposes a binary method of vehicle license plate image based on blocking - clustering. After dividing the whole image into n blocks, clustering is carried out on pixels according to their gray - scale level in each block, then binarization is performed according to the result of clustering, and all blocks＇ binarization results are combined to form the binary image of license plate image. The experiment results show that license plate features hayer excellent expression in binary image, the complexity of the background is effectively reduced at the same time, more importantly, this method is more convenient to deal with the license plates which are taken under irregular illumination. Both of them can effectively increase the real - time property and adaptability of license plate location system.

关 键 词：分块聚类 二值化 车牌图像 

分 类 号：TN915[电子电信—通信与信息系统]