多种激光传感器实验对比与评估

摘要

针对4种不同的2D(Hokuyo激光测距仪)和2.5D激光传感器(Asus xTion,Velodyne和FARO Focus),本文通过给定测量距离时改变入射角和相同入射角下改变测量距离的对照试验获得原始数据,采用随机采样一致性算法(RANSAC)进行模型拟合并计算均方差,其次用均方差来反映传感器的扫描精度,从而得出结论:除静态扫描仪FARO,其他传感器的扫描精度与入射角无关,与测量距离有关,测量距离越远,精度越低.静态扫描仪FARO的扫描精度与扫描质量设置、入射角成正比例关系,与测量距离成反比例关系.%This paper mainly focuses on four types of 2D(Hokuyo laser range finder)and 2.5D laser range sensors(Asus xTion,Velodyne and FARO Focus).Raw data was gathered with controlled trials in which we varied the incident angles and distance between the target and the sensors. Then mean square error is calculated after model fitting with the random sample consensus algorithm(RANSAC).The mean square error reflects the scanning precision of sensors.It was found that,in general,the precision of the sensors,except the FARO,is related to the measurement distance and has no relation with the incident angle.The bigger the distance is,the lower the precision tends to be.The precision of the static scanning device FARO is proportional to scanning quality configuration and incident angle and inversely proportional to measurement distance.