基于单类支持向量机的视频车辆检测研究

Video vehicle detection research based on one-class SVM classification

针对现有视频车辆检测方法存在高误检率、不完整及需要分类的样本数量大等问题,提出一种基于单类支持向量机的视频车辆检测方法。方法通过基于视频多帧统计的方法获得完整的车道背景图像,采用单类支持向量机获得车道背景模型,应用此模型检测车辆目标区域。实验结果表明,该算法具有所需样本量小,车辆检测完整,对光照的变化具有一定的自适应能力等特点。

The existing video vehicle detection methods have the defections of high false positive rate, incomplete vehicle detection and large number of samples to be classified. A video vehicle detection method based on one-class support vector machine （SVM） is proposed. Firstly, complete lane background images are obtained with video multi-frame statistics approach, and then one-class SVM is utilized to achieve lane background model. Finally, this model is applied to obtain the vehicle target area. The experimental results show that the algorithm has the characteristics of small sample size, complete vehicle detection and self-adaptation ability to illumination conditions.