摘 要：为了准确获得飞机在飞行状态下的运动数据，本文基于近景摄影测量理论、立体视觉技术，提出并实现了一种针对机身结构件轨迹、姿态、位移、变形等多种运动特性的动态视觉测量方法。主要研究以下关键技术：基于工业近景摄影测量的多相机快速自标定方法；飞行状态下相机动态定位及消除抖动技术；刚性结构件的运动轨迹及姿态的快速求取；通过多相机(>=3)协作实现非编码标志点阵列的精确匹配。在模拟飞行环境下的实验结果表明，该系统的运动轨迹姿态测量精度可达0.01mm/0.25m，关键点位移变形测量精度可达0.01mm/0.25m。基本满足飞机测试行业的精度和可靠性测量标准。

关键词：相机自标定；相机动态定位；轨迹姿态求取；多相机匹配

An optical system for the dynamic measurement of the movement characteristics of aircraft structural parts

Abstract: In order to accurately obtain the movement data of the aircraft in flight, an optical and dynamic measurement method, which can acquire the trajectory, attitude, displacement, deformation and other kinds of movement data of aircraft structural parts, is proposed and implemented based on the theory of close-range photogrammetry and stereo vision technology. The main contents include: Multi-camera self-calibration technique based on close-range photogrammetry; Camera dynamic positioning and jitter eliminating in flight technology; Rapid acquisition technique for the trajectory and attitude of rigid structure; Multi-camera (>=3) collaboration to achieve the matching of non-coding landmarks array. The simulation results show that the trajectory and attitude measurement accuracy of the system can be up to 0.01mm/0.25m, the key point displacement and deformation measurement accuracy is 0.01mm/0.25m. It suggests that the system have met the standards of aircraft testing industry measurement accuracy and reliability.
Key words: Camera self-calibration; Camera dynamic positioning; Trajectory and attitude acquisition; Multi-camera matching