多尺度远程夜间行人识别辅助驾驶研究  

Multi-scale Remote Nighttime Pedestrian Recognition Assisted Driving Research Based on Dual-threshold

作　　者：张建军[1] 杜莉[1] ZHANG Jian-jun DU Li(Institute of Electrical and Information Engineering, Beijing Polytechnic College, Beijing 100042, China)

机构地区：[1]北京工业职业技术学院电气与信息工程学院,北京100042

出　　处：《控制工程》2017年第10期2130-2136,共7页Control Engineering of China

基　　金：北京市科技计划面上项目(KM201510853004)

摘　　要：为了很好的减少夜间开车所造成的交通事故,给能见度比较差的环境下运行的车辆提供自动安全体系,设计了以远红外传感技术为基础的夜间汽车驾驶辅助体系。该技术采用远红外传感技术,获取起初信息,同时运用灰度统计方法取得感兴趣的领域;在建立多尺度概率模型的前提下,采用多帧校验系统处理方法对以上的设计进行检测和查漏。实验得出,该方案的概率模板在配合度上有了很大的改进,并且适合在市区和郊区两种环境下运用,在郊区环境下的检测率高达94%,在市区环境下的检测率是高达90%,因此该设计方案拥有很好的可行性。In order to reduce traffic accidents caused by night driving, an automatic security system for the vehicle running under poor visibility environment is provided, the night car driving assistance system based on infrared sensing technology is designed in the paper. The technology uses far infrared sensing technology to obtain the initial information, while uses the gray statistical methods to obtain areas of interest. Under the premise of the establishment of multi-scale probability model, the multi-frame check system processing method is utilized to detect and leak the above design. Experimental results show that the degree of cooperation in the probability template of the program has been greatly improved and is suitable for using in urban and suburban two environments, the detection rate is up to 94 %in the suburban environment, and as high as 90 % in the urban environment, so that the design has a very good feasibility.

关 键 词：远红外传感技术 夜间驾驶辅助 多尺度 多帧校验 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]