Metrology for Sub-Millimetric Deformation Measurement in Vacuum Environments¶
Method 1: Digital Image Correlation (DIC)¶
Digital Image Correlation (DIC), specifically 3D Stereo-DIC, is the primary candidate for full-field deformation measurement. It operates by tracking the motion of random texture patterns (speckles) on the surface of the object.
Principles and Capabilities¶
DIC compares a "reference image" (time \(t_0\)) with a series of "deformed images" (time \(t_1, t_2,...\)). The surface of the metal plate must be covered in a stochastic contrast pattern (e.g., spray paint, laser-etched texture, or natural surface roughness). Usually the algorithm divides the image into small "subsets" (e.g., \(21 \times 21\) pixels) and tracks the movement of these subsets to sub-pixel accuracy.
- Output: Full-field displacement vectors (\(u, v, w\)) and strain tensors (\(\epsilon_{xx}, \epsilon_{yy}, \epsilon_{xy}\)) for thousands of points across the plate.
- Resolution: Modern algorithms can resolve displacements of \(0.01\) pixels. At a 3-meter distance with appropriate telephoto lenses, this translates to physical resolutions of roughly \(5\) to \(20\) micrometers, well within the sub-millimetric requirement.
DIC is robust and highly adaptable to the constraints of the user's vacuum setup, provided specific architectural choices are made.
Handling the "Unaccessible" Nature: The inability to place a calibration target inside the chamber is a major hurdle for standard DIC, which relies on imaging a checkerboard in the measurement volume to determine camera positions. However, advanced DIC systems employ a "Split Calibration" or "External Orientation" workflow :
- Intrinsic Calibration: The cameras and lenses are calibrated outside the chamber using a standard target to determine focal lengths and lens distortions.
- Extrinsic Calibration: The relative orientation of the two cameras (looking through separate windows) is determined using "tie points"—natural features on the metal plate itself or a projected pattern.
- Refraction Modeling: The window parameters (thickness, index) are manually entered into the software, which then analytically corrects the light paths during 3D reconstruction. This eliminates the need for an internal calibration grid.
Handling the "Bending in Time":
DIC is inherently a snapshot technique. It captures the entire deformation field at a single instance.
- Static Bending: If the plate bends slowly (thermal expansion), cameras can capture images at low frame rates (e.g., 1 Hz) over hours, providing a stable history of deformation without the integration drift seen in accelerometers or vibrometers.
- Dynamic Bending: If the plate vibrates, high-speed cameras can be used. Modern DIC systems can operate at frame rates from 100 Hz to 100,000 Hz, allowing for the characterization of transient events or modal analysis.
Limitations and Risk Factors¶
- Speckle Pattern Quality: The technique requires a high-contrast pattern on the plate. If the plate is bare metal inside a vacuum chamber, it may be too reflective (specular) or too uniform. Without a pattern, DIC fails. The user may need to rely on "natural texture" or use a laser speckle projector (though laser speckle can decorrelate due to surface tilt).
- Lighting at 3 Meters: The inverse-square law means that lighting the object from 3 meters away requires high power. Since the chamber is a vacuum, there is no air to scatter light, which helps, but heat generation is a concern. High-intensity LED spots placed outside the viewports are the standard solution.
Theoretical Framework of Long-Working-Distance (LWD) Metrology¶
To understand the feasibility and limitations of measuring microscopic deformations from 3 meters away, we must first establish the optical and physical laws that govern remote sensing. The measurement chain is influenced by three primary factors: the spatial resolution limit imposed by diffraction and sensor discretization, the geometric dilution of precision (GDOP) caused by the observation angle, and the refractive distortion introduced by the vacuum chamber windows.
COMPARISON¶
GOM ARAMIS Adjustable (Zeiss)¶
The ARAMIS system is widely regarded as the gold standard for industrial optical metrology.
- Key Feature: Modular Hardware. The "Adjustable" series allows the user to mount the cameras on completely separate stands. This is vital for the "multi-window" setup, where Camera 1 might be 1 or 2 meters away from Camera 2.
- Key Feature: Sensor Resolution. New sensors (up to 24 Megapixels) provide high spatial resolution, which is critical for maintaining pixel density (\(<0.1\) mm/pixel) at a 3-meter distance.
- Key Feature: Blue Light Technology. GOM uses narrowband blue light illumination. This allows the use of bandpass filters on the lenses, blocking out ambient light and heat radiation from the chamber (if the plate is hot), improving signal-to-noise ratio.
- Software Capabilities: The GOM software suite includes specific workflows for "testing in climate chambers," including refraction correction modules.
Correlated Solutions VIC-3D¶
The VIC-3D system is a highly flexible, research-grade solution that is particularly strong in custom algorithm implementation.
- Key Feature: Variable Ray Origin Calibration. VIC-3D's software is renowned for its "External Orientation" calibration capabilities. It allows the user to calibrate the cameras in the lab (intrinsic) and then simply "point and shoot" at the plate through the windows to determine the stereo geometry, making it ideal for the "unaccessible" chamber constraint.
- Key Feature: Refraction Correction. Correlated Solutions has published extensive application notes and algorithms specifically for correcting measurements through thick glass viewports.
- Cost/Flexibility: Generally more open to third-party cameras and lenses than the GOM ecosystem, which allows for easier integration of ultra-long telephoto lenses (e.g., 300mm or 400mm Canon lenses) if the 3-meter distance requires extreme magnification.
Purchase the GOM ARAMIS Adjustable system (or equivalent modular VIC-3D setup).
- Configuration: Two high-resolution cameras (12MP+), two heavy-duty tripods, and two high-quality telephoto lenses (100mm - 180mm focal length).
- Justification: It offers the best balance of static/dynamic capability, inherent 3D shape measurement, and mature software for handling the vacuum window optical distortions.
Implementation Plan¶
To ensure the success of this measurement, a strict procedural workflow must be followed. This plan addresses the identified risks of calibration, refraction, and stability.
https://colab.research.google.com/drive/1WyC20wR1C-bwf1UcmWA6g4D2RWDafSzC#scrollTo=SitxV78jLJhw
Optical Setup and Geometry¶
- Distributed Sensor Placement: Do not attempt to look through a single window. Place Camera Left at Window A and Camera Right at Window B.
- Goal: Achieve a stereo angle (\(\theta\)) of at least \(15^{\circ}\).
- Calculation: For a distance of 3000 mm, the cameras should be separated by roughly \(B = 2 \cdot 3000 \cdot \tan(7.5^{\circ}) \approx 800\) mm. If windows are 1 meter apart, this is perfect.
- Lens Selection: Use 135mm or 180mm prime lenses .
- Field of View Calculation: A 135mm lens on a standard 2/3" sensor (8.8mm width) at 3000mm distance yields a Field of View (FOV) of \(\approx 195\) mm. If the metal plate is larger (e.g., 500mm), use a shorter lens (e.g., 85mm or 100mm) or a larger sensor (Full Frame).
- Aperture: Stop down the aperture (e.g., f/8 or f/11) to maximize Depth of Field (DOF), ensuring the entire plate is in focus despite the tilt angles.
Speckle generation¶
When performing 3D Digital Image Correlation (DIC) at a distance of 3 meters with telephoto lenses, your primary challenge is maintaining speckle contrast and pattern density . At that range, the speckle dots must be small enough to avoid aliasing but bright enough to be seen through the vacuum viewport.
The industry "gold standard" speckle projectors for these conditions generally fall into three categories:
Laser Speckle Projectors (DOE-Based)¶
If you need extreme brightness or a "frozen" pattern that doesn't suffer from digital pixelation (common when using tele lenses), Diffractive Optical Element (DOE) laser projectors are used.
- Coherent/Lumentum High-Density Projectors: These use a laser source passed through a DOE to create a random pattern of up to 50,000 light dots . Because they are laser-based, the depth of field is theoretically infinite—meaning the speckles stay in focus even if the metal plate bends significantly toward or away from the camera.
- Speckle-Laser Modules (IR or Blue): For vacuum chamber applications, blue laser projectors (approx. 450nm) are often used because the shorter wavelength minimizes diffraction effects when passing through the thick glass of a viewport.
For a 3-meter standoff through a vacuum viewport, DOE-based (Diffractive Optical Element) laser projectors are superior to standard projectors because they provide an "infinite" depth of field and high power density. They don't project a "video" image; instead, they split a single laser beam into tens of thousands of individual, high-contrast dots.
1. Osela Random Pattern Projector (RPP)¶
Osela is the industry leader for structured light in harsh environments. Their RPP is specifically designed for 3D mapping and DIC.


- The Technology: Uses a specialized DOE to create up to 57,446 pseudo-random dots .
- Why it fits your 3m setup: * Focusable: You can manually or factory-set the focus for a 3-meter distance to ensure each dot is crisp.
- High Power: Available in high-power classes (up to Class 3B) which is necessary to overcome the 3m distance and light loss through a vacuum window.
- Wavelength Options: You can order it in 450nm (Blue) , which is ideal for viewing through glass and ignoring thermal "glow" from the metal plate.
2. Laser Components FLEXPOINT® MVspeckle¶
This is a highly modular system frequently integrated into custom DIC rigs for materials science.


- The Technology: A laser diode combined with a "random pattern" DOE. It produces a dense cloud of speckles that are mathematically optimized to prevent "overlap" (which can confuse DIC software).
- Why it fits your 3m setup:
- Small Form Factor: The module is only 19mm in diameter, making it easy to mount precisely outside a small viewport.
- Point Density: It offers different "fan angles." For a 3m distance, you would choose a narrow fan angle (e.g., 10° or 15°) to keep the speckle pattern concentrated on your metal plate rather than wasting light on the chamber walls.
3. Coherent (High-Density Speckle Modules)¶
Coherent produces the DOEs used by many other brands, but they also sell integrated modules for laboratory and aerospace testing.
- The Technology: They specialize in Stochastic DOEs . Unlike a grid, these DOEs create a truly random intensity distribution that mimics the "spray paint" look required for high-accuracy DIC.
- Why it fits your 3m setup:
- Sub-Pixel Accuracy: The dots produced have a Gaussian intensity profile, which DIC algorithms (like GOM or Correlated Solutions) use to calculate displacement with sub-pixel precision.
4. Coherent/Lumentum (Commercial 3D Sensing Modules)¶
If you are looking for an OEM-style solution to integrate into your own housing.
- The Technology: These are often the same DOEs used in high-end LIDAR and facial recognition, but scaled up for industrial power.
- Key Advantage: They are incredibly stable over time. In a "bending over time" experiment, you cannot have the pattern "drift" or "flicker," or your measurement data will be ruined.
Custom High-Lumen 4K Solutions¶
In some advanced research setups, researchers use high-end commercial 4K laser projectors (like the Epson EB-PU series or Nikon/Leica-compatible units ) because of their massive light output (7,000+ lumens).
- Why use these? At 3 meters, a standard lab projector might be too dim. A 4K laser projector allows you to project a custom, mathematically optimized "synthetic speckle" pattern that provides better sub-pixel interpolation than a random spray-paint pattern.