Abstract: A new method based on the grads feature of the shadow which showed the candidate vehicle regions and then excluded error license plate area by license plate prior knowledge was proposed． According to the type and proportion of our license plate features it will get a smaller initial partition plate area． Then Soble operator was used for edge detection． Finally it will fast get accurate positioning by mathematical morphology． The results show that the method provides a robust approach， which can effectively detect the front vehicles in complex traffic circumstances in real time．

Key words: shadow feature, license plate location, location algorithm, mathematical morphology

摘要： 提出了一种结合车底部阴影的梯度特征寻找并确定可能存在车牌区域，然后利用车牌先验知识排除不含车牌 的错误区域，再根据我国车牌类型特征与比例特征初步分割出小范围车牌区域，并利用Soble算子进行边缘检测，最 后结合数学形态学进行快速精确定位。实验结果表明，该方法能够快速，实时有效的检测出车牌所在位置。

关键词: 阴影特征, 车牌定位, 定位算法, 数学形态学