基于 HSI 颜色空间和行扫描的车牌定位算法

摘要

为实现在复杂环境下对车牌区域进行精确定位，提出一种基于 HSI 颜色空间和行扫描的车牌定位算法。根据车牌颜色特征对彩色车辆图像进行颜色分割，结合车牌几何特征牌筛选候选区域并按设定的规则标号；在经过预处理的图像中按序提取标号候选区域，利用行扫描算法分析该区域纹理特征，确定车牌上下边界；采用形态学及垂直投影法，设定阈值，精确定位车牌区域。实验结果表明该算法具有定位准确、适用性强等特点，应用前景广泛。%To locate license plate accurately in complex environment,a car license plate location algorithm based on HSI color space and line scanning was proposed.The candidate region was segmented auording to license plate color features and geomet-rical characteristic.Then the candidate region was marked by the set rules.And,the up and down boundaries of license plate were determined based on the line scanning of the candidate region in the preprocessed image.Finally,the accurate license plate area was received using morphology and vertical projection as well as by setting the threshold.The experimental results show that the algorithm has high accuracy,strong applicability and wide range of prospects.