Multi-Wavelength Speckle Illumination Digital Holography for Parasitic-Reflection Suppression and Macroscopic Topography Reconstruction

Light is a powerful tool for measuring the world. When scientists want to know the exact shape of a surface, they often use a technique called digital holography. This method works by splitting a laser beam into two paths: one bounces off the object being studied, and the other travels a separate route before they are recombined. Where the two beams meet, they create an interference pattern, a complex map of light and dark lines that encodes the height and texture of the surface with incredible precision. Because the light waves are so sensitive, this technique can detect changes in height smaller than the width of a single atom. However, this extreme sensitivity is also its greatest weakness. In any optical system, light rarely travels in a perfectly straight line. It bounces off lenses, windows, and the back of glass slides, creating unwanted reflections that mix with the signal from the object. These "parasitic" reflections act like static on a radio, creating ghostly patterns that distort the final image and make accurate measurement impossible.

A team of researchers from the University of Stuttgart and Carl Zeiss AG has developed a new way to silence this static without losing the clarity of the measurement. They combined a technique called speckle illumination with a method that uses multiple colors of light. Instead of shining a smooth, uniform beam of light onto an object, they used a special diffuser to scramble the light into a chaotic, grainy pattern known as a speckle field. By rapidly changing this pattern many times and averaging the results, they could mathematically isolate the true surface of the object while canceling out the unwanted reflections from other surfaces. This approach allows them to measure large, complex objects that sit behind glass or have rough textures, achieving a level of clarity that was previously out of reach for this type of technology.

The core of the problem lies in how light behaves when it reflects off multiple surfaces. Imagine looking at a coin through a thick piece of glass. The light bounces off the front of the glass, then the back of the glass, and finally off the coin itself. In a standard laser setup, all these reflections arrive at the camera at the same time, interfering with each other and creating a confusing mess of ripples that hide the coin's actual features. Previous attempts to fix this involved using light sources with very short coherence lengths, such as light-emitting diodes, but this limited the depth of the object that could be measured. The researchers needed a way to keep the long-range measurement capabilities of lasers while filtering out the noise from the extra reflections.

To solve this, the team used a trick based on the unique behavior of speckle patterns. When light passes through a diffuser, it creates a random, grainy texture. This texture is highly specific to the exact distance the light travels. If the light reflects off a surface that is slightly closer or further away than the main object, the speckle pattern it carries will look completely different from the pattern on the main object. The researchers set up an experiment where they recorded hundreds of holograms of an object, each time shifting the diffuser slightly to create a new, unique speckle pattern. They then used a reference measurement of a perfect mirror to calibrate each of these images. When they mathematically combined all these images, the signal from the main object, which stayed consistent relative to the reference, added up to become clear and strong. The signals from the unwanted reflections, however, changed randomly with every new pattern. Because they were out of sync with the main object, they canceled each other out, fading into the background like noise in a crowded room.

The researchers tested this method on several challenging scenarios. First, they looked at a standard microscope slide, a thin piece of glass that naturally reflects light from both its front and back surfaces. Using a smooth laser beam, the back-surface reflection created strong, distracting fringes that ruined the image. When they switched to their new speckle method and averaged 100 different patterns, these fringes vanished, leaving a clean view of the front surface. They found that the effectiveness of this cleaning depended on how "grainy" the light was made; a finer, more chaotic pattern suppressed the noise better. In their best configuration, they reduced the noise in the image by a factor of six compared to the standard method.

Next, they moved to a more complex object: a five-cent coin placed behind a coverslip. The coin has a rough, uneven surface with raised letters and designs. The coverslip added a layer of unwanted reflection that, with standard lighting, completely obscured the coin's details, turning the image into a chaotic swirl of errors. With the new speckle technique, the coin emerged clearly. The researchers measured the height of the coin's surface and found that the noise dropped from over 54 micrometers to just 5 micrometers. This meant that the fine details of the coin, which were previously hidden, could now be seen and measured with high precision. The method successfully preserved the sharp edges of the embossed numbers while removing the ghostly interference from the glass.

Finally, the team demonstrated that this technique could measure objects that span a large distance. They arranged three metal blocks at different heights, separated by gaps of about 10 millimeters, creating a total measurement range of over 20 millimeters. They placed a coverslip in front of the blocks to introduce the same kind of parasitic reflections. Using a single camera focus, they could not see all three blocks clearly at once. However, because digital holography allows for "numerical refocusing," they could computationally shift the focus to each block one by one. The speckle method ensured that when they focused on one block, the reflections from the coverslip and the other blocks were suppressed. By combining measurements taken at eight different laser wavelengths, they were able to stitch the data together into a single, continuous 3D map. The final result showed the three blocks at their exact known positions, with an error of only two micrometers, proving that the method can handle large, multi-level objects without the need for mechanical scanning.

The success of this work relies on a delicate balance. The researchers had to ensure that the speckle patterns were different enough to cancel out the noise, but similar enough to preserve the details of the object. They found that using a specific range of light patterns allowed them to suppress the unwanted reflections from surfaces that were more than a millimeter away, while keeping the image of the target surface sharp. This approach does not require moving the camera or the object, nor does it require changing the light source to a less powerful one. Instead, it uses the mathematics of averaging to separate the signal from the noise. While the process requires capturing thousands of images, which takes time, it opens the door to measuring complex, real-world objects that were previously too difficult to analyze with high-precision optical tools. The result is a robust method for seeing the true shape of things, even when they are hidden behind glass or surrounded by confusing reflections.