摘要
车标识别系统作为智能交通的一个重要组成部分,可以帮助有关部门提前发现可疑车辆。文章采用Adaboost算法训练级联分类器,通过提取方向梯度直方图(HOG)特征进行各类车标分类器的训练,再对分类器识别出的区域作进一步筛选,排除误检区域。此方法实现的车标识别系统,能够有效地识别出车标,具有实际应用价值。
As an important part of intelligent transportation,vehicle logo recognition system can help relevant departments to detect suspicious vehicles in advance.In this paper,AdaBoost algorithm is used to train cascaded classifiers,and gradient direction histogram(HOG)features are extracted to train all kinds of vehicle logo classifiers,and then the regions identified by the classifiers are further screened to eliminate the false detection areas.The vehicle logo recognition system realized by this method can effectively identify the vehicle logo and has practical application value.
作者
叶玉双
杨洁
Ye Yushuang;Yang Jie(Jiyang College of Zhejiang A&F University,Zhuji,Zhejiang 311800,China)
出处
《计算机时代》
2020年第12期6-9,13,共5页
Computer Era
基金
浙江省自然科学基金探索项目(Q LQ20F020004)
浙江农林大学暨阳学院人才启动项目(JY2018RC04)。
