宽基线立体视觉远距离目标位姿测量  

Wide-baseline Stereo Vision for Long-range Target Position Measurement

作　　者：刘桐伯 景文博[1] 熊泽宇 吴雪妮 夏璇 LIU Tongbo;JING Wenbo;XIONG Zeyu;WU Xueni;XIA Xuan(School of Opto-Electronic Engineering,Changchun University of Science and Technology,Changchun 130022)

机构地区：[1]长春理工大学光电工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2024年第3期37-44,共8页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：高等学校学科创新引智计划(111计划D21009)。

摘　　要：立体视觉测量作为被动探测技术的一种,广泛应用于位姿测量领域。针对传统立体视觉在远距离环境下目标位姿测量受限的问题,提出一种宽基线立体视觉远距离目标位姿测量方法。基于三点定面测量原理,构建位姿测量数学模型,采用相机参数解耦策略,对测量系统所需参数进行标定。设计远距离场景下易于识别的合作目标,利用两台相机对目标进行成像,通过计算合作目标中心点的三维空间坐标,完成远距离环境下目标位姿的解算。实验结果表明,在75~85 m距离下,该方法对目标间相对位姿测量精度优于0.5%,X和Y方向均方根误差优于7.3 mm,Z方向均方根误差优于22.6 mm,可有效适用于远距离下的目标位姿测量应用。Stereo vision measurement,as a kind of passive detection technology,is widely used in the field of position measurement.Aiming at the problem that traditional stereo vision is limited in target position measurement in long-distance environment,a wide-baseline stereo vision long-distance target position measurement method is proposed.Based on the principle of three-point fixed-plane measurement,the mathematical model of position measurement is constructed,and the decoupling strategy of camera parameters is adopted to calibrate the required parameters of the measurement system;coop-erative targets that are easy to be identified in the long-distance scene are designed,and the targets are imaged with two cameras,and the resolution of target position under the long-distance environment is accomplished by calculating the three-dimensional spatial coordinates of the centroid of the cooperative targets.The experimental results show that at a distance of 75~85 m,the accuracy of this method is better than 0.5%,the root-mean-square(RMS)error in the X and Y directions is better than 7.3 mm,and the RMS error in the Z direction is better than 22.6 mm,which is effectively applicable to the application of target position measurement at long distances.

关 键 词：立体视觉 宽基线 远距离 图像处理 位姿测量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]