基于主元证据理论在车辆识别中的应用

Vehicle Identification Based on Principal Component Evidence Reasoning

摘要:

车辆识别技术作为智能交通管理系统中的研究热点和难点;在车辆识别技术中,应用Dempster- Shafer证据组合规则融合冲突信息时会产生不合理的结果;基于修正证据源的思想,提出了一种新的权重系数确定方法,该方法从证据主元角度分析,确定各组证据主元,利用该主元求出证据相容度、可信度,进而确定证据权重系数；通过新的证据冲突衡量方法,确定冲突值,归一化权重,修正证据源,按ER规则融合各组证据对目标进行识别;仿真部分以实际路面车辆车型识别为算例,将该方法与其他方法对比,结果表明:该方法能更有效地融合高度冲突的证据,减小计算复杂度,目标识别的准确性提高20%。

Abstract:

Vehicle recognition has become a hotpot in the researches on intelligent traffic management system. In the vehicle identification, using the combination rule of Dempster-Shafer evidence to fuse the collision information, the irregular results will be occurred. A new calculation method of weight coefficient is discussed in this thesis. In this method, the evidence principal elements for each group can be determined by analyzing the principal component of evidence. Then the evidence compatibility and credibility are obtained. Further, the weight coefficient of evidence can be determined through all above parameters.This thesis puts forwards a measure method for conflict of evidence This method can calculate the conflict value, then normalize weight, modify the evidence source. Further, ER evidence is used to identify the target. In the simulation progress, the example of the recognition of the vehicle type on realistic road can be utilized to compare this method with others. It can be proved that this method is more effective than others on fusing the evidence with high conflict and reducing the computational complexity. The accuracy of target recognition is improved by 20%.