碰撞预判系统中车辆宽度和位置的测量方法  

Vehicle Width and Position Measurement Method for Pre-crash System

作　　者：罗逍[1] 姚远 李沛雨[1] 杨帆[1] 张金换[1] 

机构地区：[1]清华大学,汽车安全与节能国家重点实验室,北京100084 [2]汽车噪声振动和安全技术国家重点实验室,重庆401120

出　　处：《汽车工程》2016年第2期181-184,215,共5页Automotive Engineering

摘　　要：为进一步提高碰撞预判系统的判断精度,提出了一种基于雷达和图像融合的车辆宽度和位置测量方法。首先,确定雷达坐标系与图像坐标系之间的转换关系;然后根据图像中目标的阴影或雷达探测目标在图像中的投影生成目标假设区域;通过Haar特征车辆级联分类器对生成的目标假设区域进行验证;最后对验证过的目标进行宽度和位置测量。试验结果表明,该方法能较准确地测量车辆宽度和位置,为碰撞预判系统提供重要参数。To enhance the judgment accuracy of pre-crash system,a measurement method of vehicle width and position based on radar and image fusion is proposed. Firstly,the transformation relationship between radar coordinate and image coordinate is determined; Then,the object hypothesis region is generated based on the shadow of vehicle object in the image or the projection of vehicle object in the image detected by radar,and the objective hypothesis region generated is verified by Haar-like feature classifier; Finally,the width and position of vehicle objects verified are measured. Test results show that the method proposed can accurately measure the width and position of vehicle,providing important parameters for Pre-crash system.

关 键 词：汽车安全 碰撞预判系统 车辆识别 宽度和位置测量 

分 类 号：U463.6[机械工程—车辆工程]