基于闭运算和背景减法的车牌字符分割算法

Segmenting Characters of License Plate Based on Closing Operation and Background Subtraction

提出一种基于闭运算和背景减法预处理的二值化分割算法。该方法能够削弱上下边框,减少字符与上下边框之间的粘连,同时能减少背景噪声。在此基础上利用车牌字符的垂直投影特征和先验知识处理字符的粘连与断裂,有效提高了字符分割效率。实验结果表明,该方法具有良好的适应性。

A preprocessing algorithm for binary segmentation is proposed, which is based on closing operation an background subtraction. By subtracting the image after closing operations with the original one, the contours of border are faded. Adhesion between borders and characters is reduced, while the background noise declines greatly. Then bridging the discontinuity among character lines by using vertical projection features and the prior knowledge of license plate, the efficiency of character segmentation is improved prominently. Experimental results demonstrate that the method proposed has good adaptability.