In the realm of scientific exploration, some questions are not about "whether we can see," but "whether we can see fast enough."

When metal powder melts and splatters under the intense heat of a laser, the entire process unfolds in milliseconds. When a plasma discharge completes in microseconds, spanning just 0.1 millimeters in space. When synchrotron radiation penetrates matter at near-light speed, and chemical states transform in an instant—conventional optical equipment is powerless. Industrial cameras are the key that unlocks the door to the "ultra-fast world."

From additive manufacturing laboratories in Belgium to synchrotron facilities in Japan, from plasma research institutes in Germany to microplastics detection platforms in Switzerland, industrial cameras are playing an indispensable role in the world's leading research institutions. They are not merely "cameras that take pictures"—they are "time microscopes" that allow scientists to freeze ultra-fast phenomena and reveal microscopic mechanisms.

I. Freezing the Splatter: The Additive Manufacturing Revolution at 20,000 fps

Laser Powder Bed Fusion (LPBF) is one of the core technologies in metal additive manufacturing, yet it has long been plagued by an "invisible enemy": most defects originate beneath the layer being printed, rendering conventional monitoring methods completely blind.

A research team at KU Leuven in Belgium changed this. They mounted a high-speed camera at a 25° angle outside the printing platform, capturing the spattering behavior of metal droplets during laser melting at 20,000 frames per second. Oxidized during flight, these droplets fall back onto the build surface and become inclusions, ultimately triggering a cascade of defects—microstructural anomalies, increased surface roughness, and porosity.

The research team extracted five physical features from the high-speed video stream: melt pool aspect ratio, melt pool area, mean melt pool intensity, spatter velocity, and spatter count. The key finding: a strong correlation exists between real-time spatter counts and the number of pores later confirmed by X-ray CT scanning. This means that optical signals alone can predict the internal quality of final parts during the printing process.

The breakthrough hinged on the camera's ability to resolve individual spatter events at approximately 100 μm/pixel resolution within a 120×120 pixel field of view, under 30-microsecond exposure times. Without sufficient speed and sensitivity, none of this would be possible.

II. Freezing Time: 1-Microsecond Exposure and 5,000 fps Synchrotron Imaging

If the KU Leuven case demonstrates "how fast an industrial camera can run," then the collaboration between Tohoku University and Nishikyushu University in Japan demonstrates "how short an instant an industrial camera can freeze."

The research team developed a lens-coupled high-speed X-ray imaging system called "Hayaka" (meaning "very fast" in a Japanese dialect) for synchrotron radiation 4D computed tomography and X-ray absorption fine structure imaging. At the heart of the system is a high-speed CMOS industrial camera equipped with a global shutter, achieving 3,660 fps at full 1.1-megapixel resolution and 5,000 fps at an 800×400 pixel ROI.

Even more remarkable is its temporal resolution: a minimum exposure time of just 1 microsecond. At 10-microsecond exposure, the system achieves 77-micrometer spatial resolution. This means researchers can "freeze" millisecond- or even microsecond-scale rapid structural changes and chemical state transitions.

In X-ray absorption fine structure mode, the system acquires complete spectral image data in just 0.5 seconds by combining high-speed energy scanning. The research team used this system to capture the boiling process of Japanese noodles over 150 seconds, presenting an unprecedented 3D dynamic cooking view at 0.5-second temporal resolution. Their next goal is even more ambitious: real-time tracking of chemical state changes of multiple elements inside electrodes during battery charging and discharging.

At premier large-scale scientific facilities like synchrotron radiation sources, every X-ray beam is precious. An imaging system that can capture every data point with maximum efficiency directly determines the quality and speed of scientific output.

III. Seeing the Invisible: Multi-Camera Synchronization for 3D Plasma Capture

Plasma is one of the most difficult phenomena to study experimentally—small in scale (0.1 mm diameter, 10 mm length), highly dynamic, and changing within microseconds.

Researchers at the Leibniz Institute for Plasma Science and Technology (INP) in Greifswald, Germany, chose to meet this challenge head-on with multiple industrial cameras. Their研究对象 is the kINPen plasma jet—an atmospheric-pressure cold plasma source developed at INP that generates a highly dynamic, self-luminous plasma filament, widely used in plasma medicine and materials processing.

The critical challenge: a single camera provides only a two-dimensional view. For self-luminous, highly dynamic objects like plasma filaments, 2D images alone cannot reveal the true three-dimensional spatial structure—curvature, coiling, or lateral deflection of the discharge all remain invisible. The solution: precise synchronization of multiple cameras capturing the same plasma discharge event simultaneously from different angles.

"All cameras must operate in perfect synchronization—this is the only way to capture the same features within a very short timeframe," emphasizes Artur Wittig, a research fellow at INP. Exposure times are set between 9.35 and 30.03 microseconds. Only when every camera exposes at the same microsecond and captures the same plasma filament in the same instant does 3D reconstruction become possible.

In this research, industrial cameras are no longer mere "recording tools"—they are a core component of the experimental methodology itself. Without multi-camera synchronization and nanosecond-level timing control, experimental observation of 3D plasma structures would be simply impossible.

IV. Beyond the Core: Expanding Horizons Across Scientific Disciplines

The application of industrial cameras in scientific research extends far beyond the three domains above.

Life Sciences and Biomedical Imaging. Stoelting, a life sciences research equipment manufacturer, integrated The Imaging Source's DMK 37AUX287 USB3 Vision industrial camera into its research systems, achieving high-speed image capture at up to 539 fps in a compact form factor. Teledyne DALSA launched the Falcon4-CLHS M8200-Scientific, a 67-megapixel model specifically designed for life sciences imaging.

Fluid Dynamics and Particle Image Velocimetry (PIV). Researchers studying turbulent flows employ dual high-speed cameras for simultaneous fluorescent and Mie-scattering PIV. A three-dimensional Lagrangian particle tracking system using four high-speed cameras has been developed for pulsatile pipe flow experiments.

Combustion and Propulsion Research. High-speed cameras are indispensable in combustion science. Researchers use them to capture in-cylinder flame images under various conditions. Detonation combustion research employs cameras at frame rates of 430,000 fps (2.33 μs/frame).

Materials Science and High-Velocity Impact. Researchers at the U.S. Army Research Laboratory use ultra high-speed framing cameras to study high-velocity impact phenomena using multispectral imaging.

Biomechanics and Medical Research. High-speed X-ray characterization of bone-implant motion during falls requires camera frame rates of 8,500 fps to capture sufficient visual data.

Quantum Optics and Ultra-Low-Light Imaging. Back-illuminated sCMOS cameras designed for photon-limited applications are advancing research in quantum optics and fluorescence microscopy.

V. FAQ: Common Questions for Scientists

Selection Guide for Scientific Research

Selecting the right camera for scientific research requires a different approach than industrial inspection. While industrial applications prioritize speed, throughput, and cost, research applications demand extreme temporal resolution, synchronization precision, and experimental flexibility. The following guide is distilled from the real-world cases above and common pain points reported by research teams.

1. Selection by Research Scenario

2. Selection by Key Parameters

① Frame Rate – This is often the primary specification in research. Determine the minimum frame rate based on the Nyquist criterion relative to the phenomenon's characteristic time scale. For example, if the target event lasts 1 ms, a frame rate of at least 2,000 fps is required to resolve it. For plasma discharges (microsecond-scale), rates exceeding 100,000 fps may be necessary.

② Exposure Time – Frame rate determines "how often you take a picture," but exposure time determines "how much blur is in each picture." For ultra-fast events, exposure time must be shorter than the characteristic time of motion to avoid motion blur. The 1-microsecond exposure in synchrotron imaging is a benchmark.

③ Synchronization – For multi-camera 3D reconstruction or multi-angle observation, nanosecond-level synchronization is non-negotiable. Verify whether the camera supports hardware trigger, trigger delay consistency, and PTP (IEEE 1588) network synchronization.

④ Sensitivity and Noise – Research often involves short exposures and low light levels. Quantum efficiency (QE), read noise, and dark current are critical. Back-illuminated sCMOS sensors with thermoelectric cooling are preferred for photon-limited applications.

⑤ Interface – CoaXPress offers the highest bandwidth (up to 50 Gbit/s) for ultra-high-speed imaging. 10GigE provides a cost-effective balance of bandwidth and distance for moderate-speed research. CLHS (Camera Link HS) is used in large-scale facilities for fibre-optic long-distance transmission.

3. Research Selection Checklist

Before finalizing your camera selection for a research project, go through this checklist:

• [ ] What is the characteristic time scale of my target event? (ms/μs/ns)
• [ ] What is the minimum frame rate required to resolve it? (≥2× Nyquist frequency)
• [ ] What exposure time is needed to avoid motion blur at my target speed?
• [ ] Do I need multi-camera synchronization? If so, what precision is required? (μs/ns)
• [ ] Is global shutter mandatory? (For any moving target, the answer is yes.)
• [ ] What interface bandwidth do I need for my desired resolution × frame rate?
• [ ] Does the camera offer adequate sensitivity for my light level? (QE, read noise)
• [ ] Does the SDK support my operating system (Linux/Windows/macOS) and programming environment (Python/C++/MATLAB)?
• [ ] Is the camera compatible with my existing lab infrastructure? (lens mount, cabling, trigger system)
• [ ] Have I considered data storage and real-time processing capacity? (TB/hour)

Conclusion

Looking back at these cases from the world's leading research institutions, a clear picture emerges:

At KU Leuven's additive manufacturing laboratory, a high-speed camera running at 20,000 fps revealed the hidden link between spatter and defects. At Tohoku University's synchrotron facility, an industrial camera with 1-microsecond exposure froze chemical state transitions in an instant. At the Leibniz Institute for Plasma Science and Technology, the precise synchronization of multiple cameras made the invisible three-dimensional structure of plasma visible.

The common thread: industrial cameras are no longer "accessories"—they are integral components of scientific methodology. Their frame rate defines the limit of temporal resolution; their sensitivity defines the boundary of weak signal detection; their synchronization precision defines the feasibility of multi-dimensional observation.

On the path of scientific discovery, the clearer and faster we see, the closer we get to the truth. And industrial cameras—whether from established international brands or emerging global players—are the keys that help scientists unlock the door to the ultra-fast world.