【摘要】 根据车牌的综合特征,提出了一种新的基于边缘颜色分布的车牌定位算法。该算法抓住了车牌背景与字符具有固定颜色搭配的重要特点,利用车牌区域内特有的边缘颜色分布信息并结合车牌的纹理特征,有效地滤除了大量的背景和噪声边缘,然后利用车牌的结构特征和边缘信息,并结合形态滤波的方法,以进一步确定车牌区域。实验结果表明,该算法定位准确率高、鲁棒性好,而且适用于对复杂背景下的多车牌图像进行分割。更多还原

【Abstract】 According to plate integrative characteristic,a novel locating method based on edge color distribution is introduced.The proposed method focuses on matching plate’s background color and character color.And it utilizes the edge color distribution in a license plate combined with plate texture feature,in order to remove most of the background and noise edges effectively.Then it utilizes plate structure feature and edge information and combines morphological faltering to lo- cate the license plate correctly from the candidate regions.The proposed method is very effective to locate multiple license plates in complex background.And ex- perimental results demonstrate the great robustness and efficiency of the proposed method.更多还原