🔬 optics

Phase retrieval technique from regions of unresolvable fringe density using Phase Shifting Interferometry and Liquid Crystal on Silicon (LCOS) Spatial Light Modulator (SLM)

This paper presents a method using Phase Shifting Interferometry and a Liquid Crystal on Silicon Spatial Light Modulator to recover phase information from regions with unresolvable fringe density, such as camera corners, by employing wavefront compensation.

Original authors: Indrani Bhattacharya

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Indrani Bhattacharya

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Seeing the Invisible

Imagine you are trying to take a photo of a very bumpy, wrinkled piece of fabric using a camera. To figure out exactly how bumpy it is, you shine a special light on it that creates a pattern of stripes (like a barcode) across the surface. This is called interferometry.

Normally, you can count the stripes to measure the bumps. But what happens if the fabric is so bumpy in the corners that the stripes get squished together so tightly that your camera can't see them anymore? They just look like a blurry, gray mess. This is the problem the author, Indrani Bhattacharya, is solving.

The Problem: The "Blurry Corners"

In the world of optical measurement, there is a standard method called Phase Shifting Interferometry (PSA). Think of this like taking four photos of the striped pattern, but slightly shifting the light between each shot. By comparing these four photos, a computer can calculate the exact shape of the surface.

However, this method has a weakness: The Corners.

The Issue: In the corners of the test area, the surface often changes so quickly that the stripes become incredibly dense.
The Result: The camera's sensor is like a net with holes. If the stripes are too close together (higher than the net's mesh size), they slip right through. The camera sees a flat, gray blur instead of stripes. This is called "aliasing" or "unresolvable fringe density."
The Consequence: The computer tries to calculate the shape, fails because it can't see the stripes, and the data in the corners becomes garbage (noise).

The Solution: The "Magic Mirror" (LCOS SLM)

The author uses a special piece of technology called an LCOS Spatial Light Modulator (SLM).

The Analogy: Imagine the LCOS is a smart, programmable mirror sitting in front of the light. It can change its shape pixel-by-pixel, instantly.
How it works: Instead of just taking a picture and hoping for the best, the system uses the LCOS to "pre-shape" the light. It acts like a noise-canceling headphone for light.
- If the surface has a weird bump, the LCOS creates an "anti-bump" in the light beam to cancel it out.
- This flattens the light waves before they hit the camera.

The Process: How They Fixed It

Here is the step-by-step workflow, simplified:

The First Try (The "Blended" Mess):
They take the first set of photos. In the corners, the stripes are so dense they look like a solid gray block. The computer can't read them.
- Analogy: It's like trying to read a book where the letters are printed so small they look like a solid black line.
The "Reliability Check":
The computer looks at the photos and asks, "Can I trust this data?" It calculates a "Modulation Amplitude" (a fancy way of saying "How clear are the stripes?").
- If the stripes are blurry (low contrast), the computer marks that area as "Unreliable."
- If the stripes are clear, it marks them as "Trustworthy."
The Wavefront Compensation (The Magic Trick):
This is the core innovation. The system takes the "unreliable" data from the corners and uses the LCOS mirror to fix the light before taking the next set of photos.
- It applies an "inverse" shape to the mirror. If the surface curves up, the mirror curves down.
- This flattens the light waves hitting the camera. Suddenly, those super-dense, blurry stripes in the corners spread out and become easy to see.
The Final Calculation:
Now that the light is "flattened" by the mirror, the camera can easily see the stripes, even in the corners. The computer runs the math again, and this time, it successfully calculates the shape of the entire surface, including the tricky corners.

Why This Matters

Before: If you had a lens or a mirror with sharp edges, you could measure the middle perfectly, but the corners would be a lost cause. You'd have to throw away that data.
Now: With this "Magic Mirror" technique, you can measure the entire surface, even the parts that are too steep or dense for a normal camera to handle.

Summary in One Sentence

The author built a system that uses a programmable mirror to "smooth out" the light waves hitting a camera, allowing it to see and measure surface details in areas that were previously too blurry and dense to capture.

The "Future" Idea

The paper ends by suggesting a future upgrade: using a "Windowed Fourier Transform."

Analogy: Imagine instead of looking at the whole blurry book page at once, you put a small magnifying glass (a window) over just one tiny section, read the text there, move the glass, and read the next section. This would allow them to analyze the dense corners locally without needing the mirror to fix everything first.

1. Problem Statement

The paper addresses a critical limitation in optical metrology: the failure of standard Phase Shifting Interferometry (PSI) to retrieve accurate phase information in regions of high fringe density.

The Issue: In areas such as the corners of an interferogram or near sharp surface gradients, fringe density often exceeds the spatial resolution of the camera sensor.
Consequences: This leads to aliasing (undersampling), low fringe visibility (blended fringes appearing as flat regions), and saturation. Consequently, the standard 4-step Phase Shifting Algorithm (PSA) fails, resulting in noise, artifacts, or complete retrieval failure due to the breakdown of the assumed sinusoidal variation.
Context: While traditional methods rely on high-quality fringe patterns, real-world applications involving steep phase discontinuities or specific optical setups often generate regions where fringes are "unresolvable."

2. Methodology

The proposed solution integrates Phase Shifting Interferometry (PSI) with a Liquid Crystal on Silicon (LCOS) Spatial Light Modulator (SLM) acting as both a phase shifter and a wavefront compensator.

Experimental Setup:
- Utilizes a Twyman-Green interferometer configuration.
- The LCOS SLM is employed to introduce precise phase shifts and to modulate the wavefront dynamically.
- The setup is designed to intentionally create "blended fringes" at the camera corners to test the robustness of the retrieval method.
Phase Retrieval Process:
1. Data Acquisition: Four interferograms ( $I_1, I_2, I_3, I_4$ ) are recorded with uniform phase shifts of $0, \pi/2, \pi,$ and $3\pi/2$ .
2. Standard PSA Calculation: The wrapped phase $\phi(x,y)$ is calculated using the standard 4-step formula:
  $\phi(x, y) = \tan^{-1} \left[ \frac{I_4 - I_2}{I_1 - I_3} \right]$
3. Reliability Masking (Key Innovation):
  - The Modulation Amplitude $B(x,y)$ is computed to assess signal reliability:
    $B(x, y) = \frac{1}{2} \sqrt{(I_1 - I_3)^2 - (I_2 - I_4)^2}$
  - A reliability mask is generated based on the normalized modulation amplitude. This mask identifies pixels where the fringe signal is trustworthy (non-zero modulation) versus those where fringes are unresolved or blended.
  - The LCOS SLM and the modulation mask are combined to isolate trustworthy phase data, effectively filtering out unreliable regions caused by aliasing.
4. Phase Unwrapping: A quality-guided phase unwrapping algorithm (Herraez et al., implemented via Matlab) is applied to the masked data to recover the continuous phase map.
Wavefront Compensation:
- After initial phase measurement, the LCOS SLM is used to apply an inverse wrapped phase to the system.
- This compensates for the wavefront curvature, effectively "flattening" the fringes in the high-density regions.
- The process is iterative: measure the wavefront $\rightarrow$ apply inverse phase via LCOS $\rightarrow$ re-measure to verify uniformity.

3. Key Contributions

Novel Retrieval Technique: The paper presents a method to recover phase data specifically from regions where fringes are traditionally considered "unresolvable" due to high density or aliasing.
Dual-Role LCOS SLM: It demonstrates the LCOS SLM not just as a phase shifter, but as an active wavefront compensator that can dynamically correct optical aberrations and fringe density issues in real-time.
Reliability Masking: The integration of the modulation amplitude ( $B$ ) as a reliability indicator allows the system to distinguish between valid phase data and noise/artifacts in blended fringe regions, preventing error propagation during unwrapping.
Continuation of Research: This work advances previous findings by the author (Ref [14]), moving from basic LCOS phase shifting to solving the specific problem of corner artifacts and high-frequency fringe blending.

4. Results

Experimental Validation: The system successfully generated interferograms at gray values (0, 60, 120, 180) corresponding to the required phase shifts.
Wavefront Flattening:
- Before Compensation: Interferograms showed blended fringes and intensity variations, particularly at the corners.
- After Compensation: The application of the inverse phase via the LCOS resulted in a highly uniform intensity distribution across the active area. The "blended" regions were effectively resolved, and the wavefront appeared nearly planar.
Visualization: 3D and 2D reconstructions (Figure 5) confirmed that the residual surface features were successfully isolated after Zernike polynomial subtraction, and the compensated wavefront showed a significant reduction in error compared to the uncompensated state.

5. Significance and Future Scope

Significance: This technique enhances the robustness of optical metrology systems, allowing for accurate surface profiling and wavefront sensing even in challenging conditions (high gradients, corner artifacts) where standard PSI fails. It reduces the need for complex optical realignment by using software-controlled wavefront compensation.
Future Scope: The author suggests exploring Windowed Fourier Transform (WFT) with adaptive window sizes. This could allow for local phase extraction in high-density regions without relying solely on global phase shifting, potentially offering higher resolution in complex fringe patterns.

In summary, the paper provides a practical and effective workflow for overcoming the spatial resolution limits of cameras in interferometry by leveraging the dynamic capabilities of LCOS SLMs for both phase shifting and wavefront correction.