TL;DR
This paper presents an improved method for precisely measuring tiny displacements using Fabry-Perot (F-P) etalon interference patterns. By refining image processing techniques, specifically a novel secondary coordinate rotation (SCR) and the multichord averaging method (MCAM), the researchers achieved highly accurate center-finding for the interference rings, leading to microdisplacement measurements with an uncertainty of approximately 0.03 µm over a 5 mm range.
What Problem Does This Paper Address?
Measuring very small displacements (micro- or nanometer-scale) is crucial in many scientific and engineering fields, from precision manufacturing to fundamental physics experiments. Optical interferometry, particularly using Fabry-Perot (F-P) etalons, offers high resolution for such measurements. However, accurately determining the center of the concentric interference rings in the F-P pattern is a significant challenge. Existing methods, while functional, can be limited by factors like pixel resolution, noise, and the accuracy of circle-fitting algorithms. This paper aims to overcome these limitations by developing and validating a more robust and accurate image processing pipeline for F-P interferometry, specifically targeting microdisplacement measurements.
Methods
The core of this work lies in an improved data processing method for analyzing 2D F-P fringe patterns. The authors introduce a Secondary Coordinate Rotation (SCR) method, which builds upon a previous Coordinate Rotation (CR) method. The SCR method involves a 45-degree rotation of the coordinate system, effectively reducing the virtual pixel interval by half and improving the accuracy of calculating the center coordinates of the concentric rings.
Beyond the improved subdivision method, the paper also investigates and compares two circle-fitting algorithms for determining the ring centers:
- Circular Regression Method (CRM): A standard approach for fitting circles to data points.
- Multichord Averaging Method (MCAM): Proposed as a more suitable method for this application.
The processing pipeline involves several steps: finding approximate circle centers, extracting effective rings, calculating ring peak coordinates (using local univariate quadratic regression), selecting the optimal ring, applying the virtual pixel subdivision (SCR), accurately calculating center coordinates using either CRM or MCAM, and finally, calculating the microdisplacement.
Datasets
Not discussed in the paper. The experiments were conducted using a custom-built setup with a Hg spectral lamp, an F-P etalon, an objective lens, and an Olympus PEN-F camera.
Key Insights and Ideas
- The power of coordinate rotation for sub-pixel accuracy: The SCR method, by rotating the image data, effectively "upsamples" the information, allowing for more precise localization of the ring centers than standard methods.
- MCAM outperforms CRM for ring fitting: The study demonstrates that the Multichord Averaging Method (MCAM) provides more accurate and stable results for determining the center coordinates of the F-P interference rings compared to the traditional Circular Regression Method (CRM). This is crucial because the accuracy of displacement measurement is directly tied to the accuracy of these center coordinates.
- Focusing is critical, but the optimal ring is key: While precise focusing is important for image quality, the paper highlights that identifying the "optimal ring" (likely based on signal-to-noise ratio or fringe clarity) is paramount for accurate center determination.
- Subdivision significantly improves accuracy: Both the CR and SCR subdivision methods dramatically improve the accuracy of center coordinate calculations compared to using raw pixel data. SCR, however, offers a superior improvement.
Key Results
- Improved center coordinate accuracy: The SCR method, when combined with MCAM, yields higher accuracy in determining the center coordinates of the interference rings.
- Standard deviation of center: The standard experimental deviation of the circle center is approximately 0.009 µm.
- Displacement uncertainty: The extended uncertainty of the microdisplacement measurement over a 5 mm range is approximately 0.03 µm (with a coverage factor k=2).
- Linearity: The micro-range displacement measurements showed excellent linearity.
- Effect of N": The standard deviation of the center coordinates decreases as the number of line segments (N") used in the fitting process increases, with diminishing returns beyond N"=50.
Limitations and Future Work
- Environmental dependence: The authors acknowledge that the achieved accuracy is dependent on the laboratory environment and the specific parameters of the experimental equipment. They suggest that improvements in these areas could lead to even better results.
- Pixel size: The paper notes that if the average pixel interval (w) of the area array detector were to shrink (e.g., to w/2), the standard deviation of the circle center could potentially be halved. This points to the importance of detector resolution for ultimate accuracy.
- Future work: The authors believe their processing method can be extended to other F-P interferometry applications and other area-array detectors, suggesting broad applicability. They also imply that further optimization of experimental conditions and equipment could push the accuracy limits even further.
Summary updated