基于YUV空间和边缘能量的车牌检测方法  

License plate detection method based on YUV space and edge energy

作　　者：张晶[1] 高彦彦[1] 史源平[1] 

机构地区：[1]石家庄学院机电学院,河北石家庄050035

出　　处：《电视技术》2017年第9期171-175,共5页Video Engineering

基　　金：河北省科学技术厅项目(15270349);河北省高等学校科学研究计划项目(ZC2016107);石家庄学院应用型课程开发与建设项目(YYKC-201614);河北省教育科学十二五规划课题(140650)

摘　　要：针对复杂背景,提出了一种新的方法来检测车辆牌照。首先,将图像由RGB颜色空间转换到YUV颜色空间,因为车牌通常在尾灯附近一定范围内,所以利用V分量探测后车灯,再根据先验知识确定车牌的候选区域。接着,在这个区域内,利用Sobel算子获取图像垂直边缘,并采用一种新的边缘能量图方法进行处理,在垂直与水平方向分别确定两个点从而以此作为矩形的顶点来检测车牌。实验结果表明,该算法检测车牌的准确率为93%,运算时间约为0.21秒,可以克服复杂背景的影响,实现实时及准确性能。Aiming at complex background, a new method is proposed to detect vehicle license plate. Firstly, the image is trans- formed from RGB color space to YUV color space, and the rear light is detected by V component, because the license plate is usu- ally within the range of the rear light, thus the candidate area of the license plate is determined according to prior knowledge. Then, Sobel operator is used to obtain the vertical edge of the area, and a new method of edge energy map is used to determine the two points in the vertical and horizontal directions respectively, so as to locate the license plate as the vertex of the rectangle. The experimental results show that the accuracy of the proposed algorithm is 93% and the computation time is about 0.21 seconds. It can overcome the influence of complex background and achieve real-time and accurate performance.

关 键 词：RGB颜色空间 YUV颜色空间 V分量 垂直边缘 边缘能量图 矩形 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]