A Novel Vision-Based Pose Measurement Method Considering the Refraction of Light

Wei Liu, Xin Ma, Xiao Li, Yi Pan, Fuji Wang, Zhenyuan JiaSensors 2018 · DOI

TL;DR

This paper introduces a novel vision-based pose measurement method for aircraft testing in wind tunnels. By developing an imaging model that accounts for light refraction through the observation window, the researchers significantly improve measurement accuracy compared to traditional methods, achieving a 57% increase in displacement accuracy and 33.6% in angle accuracy.

What Problem Does This Paper Address?

Accurately measuring the six-degree-of-freedom (6-DOF) pose (position and orientation) of aircraft models during wind tunnel tests is crucial for understanding aerodynamic performance, especially during store separation. Traditional contact-based measurement systems struggle in the harsh wind tunnel environment (high speeds, limited space). Vision-based systems offer a non-contact, high-speed alternative. However, a major challenge arises from the need to protect cameras behind thick glass observation windows. This glass, combined with the high-speed airflow, causes light refraction, which significantly distorts images and leads to substantial measurement errors when using standard camera models. Existing vision-based methods often overlook or inadequately address this refraction effect, limiting their accuracy in real-world wind tunnel applications.

Methods

The core of this paper is a new imaging model that explicitly incorporates the effects of light refraction through the observation window. Key methodological contributions include:

  • Refraction-based Imaging Model: Unlike the traditional pinhole camera model, this new model accounts for the multi-layer media (air, glass, airflow) and their respective refractive indices. It models the light path from the target to the camera, considering the bending of light at each interface.
  • Linear Calibration of Window Normal Vector: A novel method is presented for linearly calibrating the normal vector of the observation window. This is crucial because the window's orientation significantly influences the refraction.
  • Step-by-Step Calibration: To avoid parameter coupling issues (where optimizing all parameters simultaneously leads to suboptimal results for specific ones), a step-by-step calibration approach is employed. This involves:
  • Calibrating intrinsic camera parameters without the window.
  • Determining the window's normal vector.
  • Calibrating external camera parameters and medium layer thicknesses.
  • Coplanarity Constraint Matching: A matching method based on coplanarity constraints is used to reconstruct 3D points from stereo images, specifically designed to handle the multi-layer refraction.
  • Ultra-thin Retro-reflective Markers: To maintain aerodynamic performance and provide clear features, ultra-thin (20 µm) retro-reflective markers are printed and applied to the target.

Datasets

The paper does not explicitly mention or introduce new datasets for training. Instead, it focuses on experimental validation using a controlled laboratory setup. The "dataset" used for evaluation consists of measurements taken from a calibrated electronic control platform that moves a target in known positions and orientations, allowing for comparison between the proposed method and a conventional pinhole model.

Key Insights and Ideas

  • Refraction is Not Trivial: The paper strongly demonstrates that the seemingly simple act of looking through glass in a high-speed environment introduces significant optical distortions that cannot be ignored for accurate pose estimation. The observed corner shift of up to 2 pixels due to a 1 cm thick glass window highlights this.
  • Step-by-Step Calibration is Key: The authors found that a holistic optimization for all camera and window parameters can lead to inaccurate estimations of the window's normal vector. Their sequential calibration approach effectively decouples these parameters for more robust results.
  • Ultra-thin Markers are Practical: The development of ultra-thin markers is a practical engineering solution that balances the need for distinct visual features with the aerodynamic requirements of the test models.
  • Coplanarity Constraint for Robust Matching: The proposed matching method, leveraging coplanarity constraints derived from the refraction model, is essential for accurately correlating feature points between stereo images despite the refractive distortions.

Key Results

The paper reports significant improvements in pose measurement accuracy:

  • Displacement Accuracy: Increased by approximately 57%.
  • Angle Measurement Accuracy: Increased by approximately 33.6%.
  • Mean Square Error (MSE) Reduction:
    • Pinhole Model: X, Y, Z displacement MSEs of 0.216 mm, 0.221 mm, 0.215 mm; Pitch, Yaw, Roll angle MSEs of 0.183°, 0.171°, 0.826°.
  • Refraction Model: X, Y, Z displacement MSEs of 0.091 mm, 0.092 mm, 0.090 mm; Pitch, Yaw, Roll angle MSEs of 0.090°, 0.093°, 0.548°.

These results were validated in a laboratory setting using a high-accuracy electronic control platform.

Limitations and Future Work

  • Limitations:
    • The current study was conducted in a laboratory setting, simulating wind tunnel conditions. While the results are promising, direct validation within an actual high-speed wind tunnel with varying airflow conditions would be ideal.
  • The paper acknowledges that the current method does not fully account for the dynamic effects of high-speed airflow on the glass itself (e.g., vibration or deformation), which could introduce further errors.
  • Future Work:
    • The authors explicitly state their next step is to "consider the airflow influence in a wind tunnel test with time-varying wind speed to further enhance the vision measurement accuracy." This suggests investigating dynamic airflow effects and potentially adapting the model for real-time adjustments.
  • Further exploration of the method's applicability to other complex measurement scenarios beyond wind tunnels could also be a future direction.

Summary updated

Abstract

Nowadays, due to the advantages of non-contact and high-speed, vision-based pose measurements have been widely used for aircraft performance testing in a wind tunnel. However, usually glass ports are used to protect cameras against the high-speed airflow influence, which will lead to a big measurement error. In this paper, to further improve the vision-based pose measurement accuracy, an imaging model which considers the refraction light of the observation window was proposed. In this method, a nonlinear camera calibration model considering the refraction brought by the wind tunnel observation window, was established first. What's more, a new method for the linear calibration of the normal vector of the glass observation window was presented. Then, combining with the proposed matching method based on coplanarity constraint, the six pose parameters of the falling target could be calculated. Finally, the experimental setup was established to conduct the pose measurement study in the laboratory, and the results satisfied the application requirements. Besides, experiments for verifying the vision measurement accuracy were also performed, and the results indicated that the displacement and angle measurement accuracy approximately increased by 57% and 33.6%, respectively, which showed the high accuracy of the proposed method.
BibTeX
@inproceedings{liu2018novel,
  author = {Wei Liu and Xin Ma and Xiao Li and Yi Pan and Fuji Wang and Zhenyuan Jia},
  title = {{A Novel Vision-Based Pose Measurement Method Considering the Refraction of Light}},
  booktitle = {Sensors},
  year = {2018},
  doi = {10.3390/s18124348},
}
Figure