Real-Time Pose Measurement Framework of Wind Tunnel Aircraft Models Based on a Monocular Time-of-Flight Camera

Area of Science:

• Aerospace Engineering
• Computer Vision
• Metrology

Background:

• Accurate, real-time aircraft model attitude measurement is crucial for aerodynamic analysis in wind tunnel experiments.
• Existing non-contact measurement methods face challenges in achieving high precision and real-time performance.

Purpose of the Study:

• To develop a novel pose measurement framework for high-precision, non-contact attitude acquisition of aircraft models in wind tunnels.
• To fuse keyframe global registration with non-keyframe local registration for robust and efficient pose estimation.
• To address the limitations of current methods in terms of accuracy, robustness, and computational efficiency.

Main Methods:

• A monocular Time-of-Flight (ToF) camera is employed for data acquisition.
• A novel hand-crafted local feature combined with TCF-RANSAC is used for robust keyframe global registration.
• The GICP algorithm is utilized for fast, incremental local registration of non-keyframes, ensuring continuity and avoiding local optima.

Main Results:

• The proposed framework achieves a single-axis rotation angle error of less than 0.03 degrees.
• Processing speeds exceed 40 FPS, meeting real-time measurement demands.
• Demonstrated superior accuracy and robustness compared to existing methods, reducing rotation angle errors by 9% to 39%.

Conclusions:

• The developed framework provides a high-precision, real-time solution for aircraft model attitude measurement in wind tunnel experiments.
• The study identifies an 'axis-sensitivity' in monocular ToF pose estimation, with lower errors around the optical axis.
• Findings offer practical guidance for camera placement and system design in aerodynamic measurement applications.