一种基于RCNN的翻车事故检测算法  

A vehicle rollover accident detection algorithm based on RCNN

作　　者：刘杰[1] Liu Jie(School of Computer, GuangXi Economic Management Cadre College, Nanning, Guangxi 530007, China)

机构地区：[1]广西经济管理干部学院计算机系,广西南宁530007

出　　处：《计算机时代》2019年第1期55-58,共4页Computer Era

基　　金：2018年度广西高校中青年教师基础能力提升项目(2018KY1272);2015年国家社会科学基金项目(15XTQ010);2015年广西高校科学技术研究项目(KY2015YB351)

摘　　要：为了研究复杂环境下交通事故侧翻车辆的检测识别问题,采用RCNN检测算法,对一定数量的常见车辆事故场景和车身关键部位进行模型训练,并在此基础上结合空间辅助定位模型加强车辆侧翻检测分类。实验结果表明,该组合方法有效地提升了翻车事故的检测识别水平,并对误检率和漏检率有显著的抑制作用。In order to study the detection and identification of traffic accident rollover vehicles in complex environments, the RCNN detection algorithm is used to train a model with certain number of common vehicle accident scenes and key parts of the vehicle body. On this basis, the detection of vehicle rollover is enhanced with the space-aided positioning model. The experiment result shows that the combined method effectively improves the detection and recognition level of the vehicle rollover accident and has a significant inhibitory effect on the false detection rate and the missed detection rate.

关 键 词：RCNN 空间辅助定位 翻车事故检测 交通安全 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]