基于改进特征提取的BP神经网络车牌识别

Recognition of vehicles' license plates by BP neural network based on improved feature-extraction

车辆牌照的自动识别是计算机视觉与模式识别技术在智能交通领域的重要应用。为提高车牌字符识别系统对字符的整体识别能力,对车牌上的汉字与数字字符做分类处理,利用粗网格、边框和笔划密度方法更好地提取识别对象特征,最后采用BP神经网络进行字符识别。实验表明,本文所采用的方法能达到较好的识别效果,说明多种特征提取方法和神经网络识别技术的有机结合能显著提高系统性能。

According to the structure of automotive suspension system, in the paper the model of the second order step and the frequency response were analyzed. Parameters were computed for each prepared case. The paper chose the best design based on the comparison of the step and the frequency response among different coefficient of elasticity and damp constant by using the simulation of MathCAD.