摘要
提出了一种基于特征尺寸成像变化规律的单目视觉测距算法,采用了单目相机和加速度计的测距方案,通过建立相机-目标距离与目标特征尺寸成像的关系,利用目标特征尺寸成像相对变化量的统计特性及特征尺寸与相机-目标距离的比值关系对距离输出进行估计并逐步修正,有效提高了单目测距精度。测距试验结果表明:即使在没有固定参照物的情况下,该方法测距精度依然优于传统单目测距方法。
This paper proposed a monocular vision ranging algorithm based on the variation law of feature size imaging.The distance measurement scheme of monocular camera and accelerometer was adopted.By establishing the relationship between camera-target distance and target feature size imaging,the distance characteristic was estimated and corrected step by step using the statistical characteristics of the target feature size imaging relative variation and the ratio of the feature size to the camera-target distance and the accuracy of monocular ranging could be improved effectively.The distance test results show that this method is superior to the traditional monocular ranging method even in the absence of a fixed reference object.
作者
向召利
马宝成
XIANG Zhaoli;MA Baocheng(Sichuan Aerospace Electronics Research Institute,Chengdu 610100,China)
出处
《兵器装备工程学报》
CAS
北大核心
2020年第2期148-151,共4页
Journal of Ordnance Equipment Engineering
关键词
单目
测距
特征尺寸
统计特性
monocular
ranging
feature size
statistical characteristic