TL;DR
This paper introduces a novel vision-based system for precisely measuring the 6-DOF pose (position and orientation) of high-speed, rolling targets within a supersonic wind tunnel. By employing ultrathin retro-reflective markers, a spatial coding marker layout, and a robust axis-fitting method, the system achieves high accuracy and overcomes challenges like marker occlusion and aerodynamic interference.
What Problem Does This Paper Address?
Accurate pose measurement of aerodynamic models in wind tunnels is crucial for aircraft design and validation. However, existing methods struggle with the demands of high-speed, rolling targets, especially in the confined and often turbulent environment of a wind tunnel. Challenges include:
- High Speed and Rolling Motion: Traditional measurement techniques can't keep up with the rapid motion and complex rotations, leading to lost data or inaccurate readings.
- Aerodynamic Interference: Markers must be small and thin to avoid altering the airflow around the model, but this makes them difficult to detect and track.
- Marker Occlusion: As the target rolls, markers can become hidden from view, leading to incomplete pose information.
- Environmental Constraints: Wind tunnels are often dark, and observation windows can be small, further complicating image acquisition.
This paper aims to bridge this gap by developing a non-contact, high-precision vision system specifically designed for these challenging conditions.
Methods
The proposed system integrates several key innovations:
- Ultrathin Retro-Reflection Markers: To address the aerodynamic constraint, the researchers developed ultrathin (20 µm) retro-reflective markers using a rolling-printed method. This ensures minimal impact on the model's shape while providing high signal-to-noise ratio (SNR) images, even at high frame rates.
- Spatial Coding Marker Layout: To combat marker occlusion and ensure accurate identification, a novel marker layout based on spatial coding is introduced. Markers are grouped and their relative positions encoded, allowing for robust identification even if some markers are lost during rolling. This method uses distance ratios between markers within a group to uniquely identify each group.
- High-Speed Image Acquisition: The system utilizes high-speed cameras (up to 5000 fps) to capture the rapid motion of the targets.
- Binocular Vision: Two cameras are employed to reconstruct the 3D coordinates of the markers.
- Non-Linear Imaging Model: A non-linear imaging model is used to account for lens aberrations, improving projection accuracy.
- Case Deletion Diagnostics for Axis Fitting: To accurately determine the target's central axis, a robust axis-fitting method based on Case Deletion Diagnostics is employed. This technique effectively identifies and removes outlier points caused by mismatched markers, enhancing the stability and precision of the axis estimation.
- Centroid Position Calculation: The centroid position is optimized using a multi-point approach with weighted markers, considering variations in marker brightness and circularity.
- Absolute Pose Solution: The system establishes an initial reference coordinate system (IRCS) and a local coordinate system (LRCS) to transform marker coordinates and derive the target's absolute pose.
Datasets
Not discussed in the paper. The experiments were conducted using custom-built experimental setups in a laboratory and a supersonic wind tunnel.
Key Insights and Ideas
- The "Rolling-Printed" Marker Fabrication: This technique is a clever solution to the conflicting demands of aerodynamic smoothness and marker visibility. It allows for markers that are both thin enough not to disturb airflow and reflective enough for high-speed imaging.
- Spatial Coding for Robustness: The spatial coding approach for marker layout is a significant advancement. It moves beyond simple marker identification to a more robust system that can handle partial occlusion, a common problem with rolling objects. This is crucial for maintaining continuous pose tracking.
- Case Deletion Diagnostics for Outlier Rejection: Applying this statistical method to axis fitting is a smart way to improve the reliability of the system. By systematically identifying and removing points that don't fit the expected model, it significantly reduces errors introduced by imperfect marker detection or matching.
- Integrated Approach: The paper's strength lies in its holistic approach, combining novel marker technology, a smart layout strategy, and advanced image processing and fitting algorithms to tackle a complex measurement problem.
Key Results
The experimental results demonstrate the effectiveness of the proposed method:
- Position Measurement Precision: Less than 0.16 mm.
- Pitching Angle Precision: Less than 0.132°.
- Yaw Angle Precision: Less than 0.132°.
- Roll Angle Precision: 0.712°. (The authors note this is primarily due to the small diameter of the tested object).
The system was successfully tested in both laboratory settings and a supersonic wind tunnel, showing high accuracy and stability for high-speed, rolling targets.
Limitations and Future Work
- Limitations:
- The paper acknowledges that factors like the large measuring range and camera resolution limitations can affect pose measurement precision.
- Marker extraction accuracy can be impacted by small marker size, surface profile variations, and deformation during rolling.
- The roll angle precision is noted to be influenced by the small diameter of the tested object.
- Future Work:
- The authors state their future research will focus on improving pose measurement precision through algorithm optimization.
Summary updated