提出了一种汽车牌照的自动识别算法:先对汽车图象进行直方图均衡、平滑去噪、二值化

预处理,再根据车牌上文字变化特点快速、准确地从复杂背景中分割车牌,最后采用投影-变换系

数法提取汉字、英文和阿拉伯数字的不同维数的特征,送到相应的BP神经网络中进行识别,较好

地解决了汽车牌照的自动识别问题。

-A vehicle license automatic recognition algorithm: first automotive images histogram equalization, smoothing denoising, Binarization pretreatment, and then plate on the basis of changes in the characteristics of text quickly and accurately from a complex background of partition plates Finally, the use of projection- transform coefficients extracted Chinese characters, English and Arabic numerals in the different dimension of features, corresponding to the BP neural network to identify, can solve the automatic recognition of vehicle license issues.