面向车辆牌照的L快速二值比算法

Algorithm of L Quick Binarization Catering to License Plate of Vehicle

就车辆牌照灰度图像的二值化问题 ,比较 Otsu算法和全局动态阈值算法 ,讨论将字符从车牌背景中分割出来的最佳阈值选取问题 .从汉字字符识别角度出发 ,通过引入光照强度因子 L,对全局动态阈值算法进行改进 ,提出 L快速二值化算法 .实验结果表明 ,L快速二值化算法的处理效果 ,优于 Otsu算法和全局动态阈值算法 .二值化处理后的车辆牌照字符笔划清晰、饱满、无断裂 ,克服了光照因素的影响 ,更有利于汉字字符的识别 .

With regard to the binarization of the greyscale image of vehicle license plate, a comparison is made between Otsu algorithm and algorithm of globally dynamic threshold; and a discussion is held on the choice of optimal threshold value separating character from license plate background. Starting from Chinese character recognition, the algorithm of globally dynamic threshold is improved by leading into it factor L of illuminance; and the algorithm of L quick binarization is advanced. As indicated by experimental results, algorithm of L quick binarization has a treating effect better than those of Otsu algorithm and algorithm of globally dynamic threshold. After treatment of binarization, the character on license plate of vehicle shows clear, full and unbroken strokes by overcoming the effect of illumination factor, which is more advantageous to Chinese character recognition.