摘要
飞机操纵系统输入是飞行性能、安全监测的重要数据来源,针对传统飞行数据采集方式需对真机进行改装并加装传感器所带来的改造成本大且影响飞行员操纵感觉的弊端,提出一种基于视觉的飞机操纵系统输入端位置测量技术。为避免视觉输入中飞机驾驶舱复杂环境的影响,将随机生成的二维码固定在飞机操纵系统输入端,使用计算机视觉技术识别二维码并进行特征点提取,通过PnP(Perspective-n-Point)问题的求解将飞机操纵输入端的位移测量转换为相机位姿求解。实验结果证明了所提方法的有效性,并通过与红外测距仪的对比实验验证了本文算法在精度上的优势,其测量平均误差在2 mm内,能达到实时处理速度,丰富了飞行数据的采集方式。
The input of aircraft control system is an important data source for flight performance and safety monitoring.In view of the disadvantages of the traditional flight data acquisition method that requires the modification of the real aircraft and installs sensors, which is costly and affects the pilot’s feeling of handling, a vision-based technology for measuring the position of the input terminal of aircraft control system is proposed.In order to avoid the influence of the complex environment of aircraft cockpit in the visual input, the randomly generated QR code is fixed at the input terminal of the aircraft control system.The computer vision technology is used to identify QR code and extract the feature points.Through the solution of PnP(Perspective-n-Point) problem, the displacement measurement of the aircraft control input is transformed into the solution of camera pose.Experimental results indicate that the proposed method is feasible.Compared with infrared rangefinders, it is verify that the proposed method has higher accuracy with the average measurement error within 2 mm, and it can achieve real-time processing speed, which enriches the way of flight data acquisition.
作者
杜官明
孙瑾
高敬博
DU Guan-ming;SUN Jin;GAO Jing-bo(College of Civil Aviation,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China)
出处
《测控技术》
2023年第2期81-86,128,共7页
Measurement & Control Technology
基金
国家自然基金青年科学基金项目(61702260)。
关键词
飞机操纵系统输入端
计算机视觉
PNP
飞行数据采集
input terminal of aircraft control system
computer vision
PnP
flight data acquisition
