基于支持向量机与结构矩的车型识别实时鲁棒算法

Real-Time Robust Algorithm of Vehicle Recognition Based on Support Vector Machines and Structural Matrix

车型分类精度是目前交通环境中的研究热点.提出了一种鲁棒性好实时性强的基于支持向量机的机动车型识别方法,对车辆的类型进行分类,选择合适的结构矩为车辆特征,以支持向量机方法为学习分类器,获得了较高的车辆分类精度,在一定程度上解决了车辆分类难的问题.现场视频处理的结果表明:该法鲁棒实时有效,且车型分类精度得到了一定提高.

The research of vehicle classification precision in transportation environment is hot recently.A real-time robust vehicle recognition method is proposed based on support vector machine.The method carries on a classification of vehicle type,and takes structural matrix as the vehicle characteristics and support vector machine method as the learning classifier system.It obtains high vehicle classification precision and solves the problem of vehicle classification.Experiments result of field transportation image sequence show that the algorithm is robust and efficient,and vehicle classification precision is improved.