节点文献
基于Log-Gabor小波变换和证据推理的车型识别
Vehicle category recognition based on Log-Gabor wavelets transform and DS theory
【摘要】 针对车型识别问题,提出了一种基于特征车的车型识别方法——基于Log-Gabor小波变换和DS证据推理的车型识别算法。该算法先对特殊车辆图像进行多分辨率的Log-Gabor小波变换,最后形成车辆Log-Gabor特征。将1-a-1多分类SVM应用于基本概率分配函数的确定,使用证据推理的方法得到车型识别的结果。实验结果表明该方法是有效、可行的。
【Abstract】 To solve the problem of vehicle category recognition,this paper proposed a recognition algorithm based on Log-Gabor wavelets transform and DS theory.Used multi-scales Log-Gabor filters to transform the images of vehicles to Log-Gabor vectors and used,SVM of "1-against-1" approach to assign the basic probability numbers.Then,concluded a decision of vehicle category by model of DS theory.Experiment results prove that the introduced algorithm is available.
【关键词】 车型识别;
Log-Gabor;
支持向量机;
Dempster-Shafer证据推理;
【Key words】 vehicle category recognition; Log-Gabor; SVM(support vector machines); DS theory;
【Key words】 vehicle category recognition; Log-Gabor; SVM(support vector machines); DS theory;
【基金】 国家自然科学基金资助项目(69674012);重庆市自然科学基金资助项目(2006BA6016)
- 【文献出处】 计算机应用研究 ,Application Research of Computers , 编辑部邮箱 ,2009年03期
- 【分类号】TP391.41
- 【被引频次】6
- 【下载频次】287