基于HSV模型和改进AdaBoost算法的车牌检测

提出了一种基于HSV颜色模型和改进AdaBoost算法的车牌检测方法。针对传统AdaBoost算法在训练过程中出现的过配现象和检测率偏低问题,文中在传统AdaBoost算法的基础上对其权值更新规则和弱分类器加权参数做了改进,并通过利用HSV颜色模型构建第一层强分类器,并构建成级联分类器应用于车牌检测。实验证明使用该方法得到的车牌检测器不仅提高了车牌检测率和检测速度,并在一定程度上避免了过配现象产生

License Plate Detection Based on HSV Color Model and Improved AdaBoost Algorithm

This paper presents a method for license plate detection based on improved AdaBoost algorithm and HSV color model.With respect to the overfitting and low detection rate appeared in the training process of traditional AdaBoost algorithm,this article makes some improvements to weight updating rule of traditional AdaBoost algorithm and weight parameter of weak classifiers,and constructs the first strong classifier by HSV color model,thus a cascade classifier for license plate detection.Experimental results show that license plate detector obtained by this method improves the detection rate and detection speed of license plate,and to some extent avoids overfitting.