基于形态学梯度重建的车牌定位方法

准确定位车牌是车牌识别的重要基础。针对复杂环境下车牌图像容易受背景、光照等因素的影响而导致车牌定位精度较低的问题,提出了一种基于形态学梯度重建的车牌定位方法。该方法首先利用颜色信息确定车牌候选区域;然后利用矢量梯度算子获取候选区域中车牌图像的梯度,利用形态学梯度重建运算提取具有车牌特征的图像结构,同时抑制非车牌特征的图像结构;最后利用车牌固定的宽长比先验信息对矩形区域进行提取,最终得到准确的车牌定位结果。实验结果表明,提出的车牌定位方法能在复杂环境下快速、准确地定位车牌,且具有较高的鲁棒性和实时性。

A novel license plate location method based on morphological gradient reconstruction

Accurate license plate location is the basis of license plate recognition. The existing approaches of license plate location are easy to be affected by background and illumination, which leads to a low accuracy of license plate location. To solve this issue, we propose a novel license plate location method based on morphological gradient reconstruction. Firstly, the candidate license plate regions are detected through the color information of a license plate. Secondly, vector gradient operator is used to obtain the gradient image of the candidate areas, and then morphological gradient reconstruction operation is used to extract the image structures with the characteristic of license plate shape and suppress the image structures without the characteristic of license plate shape. Finally, the rectangle areas are extracted according to the prior information of the fixed width length ratio of a license plate. Experimental results show that the proposed method can accurately locate the license plate in complex environment in real time with high robustness.