July 8th, 2026
When plasma becomes visible

Multi-camera image processing for new insights into highly dynamic plasma jets

Plasma jets, in which ionized gas is emitted from a source in the form of a focused, highly dynamic, light-emitting structure, play a central role in numerous technological and medical applications, ranging from materials processing to plasma medicine. At the same time, plasma discharges are among the phenomena that are most difficult to study experimentally: They are small-scale, highly dynamic, erratic, and change within a few microseconds.

At the Leibniz Institute for Plasma Science and Technology e.V. (INP) in Greifswald (Germany), the “Medical Plasma Source Systems” (MPS) research group is tackling this challenge. Under the direction of Dr. Torsten Gerling, the team is investigating the fundamental properties of plasma sources used in medicine, including through the use of image-processing-based measurement methods. A current research focus is the experimental investigation of the plasma discharge of the so-called kINPen plasma jet, an ambient-pressure cold plasma source developed at the INP that generates a highly dynamic, self-luminous plasma filament. To capture the highly dynamic discharge structure in three dimensions, the researchers have relied from the outset on a synchronized multi-camera setup using industrial cameras from IDS.

The kINPen-Plasmajet under examination is an ambient-pressure cold plasma source developed entirely at the INP; the plasma it generates leaves the device as an effluent and exhibits a highly dynamic discharge structure (1 µs period) with a very small spatial extent (0.1 mm diameter, 10 mm length). This combination of rapid temporal changes and small scale makes the kINPen a suitable reference system for experimentally investigating the spatial structure and propagation of individual plasma jet discharges.

Head of the kINPen plasma source with time-resolved discharge structure (16 µs exposure time).

It is precisely this spatial dimension that is the focus of the current works: “Our focus is on the three-dimensional structure of the plasma discharge,” explains Artur Wittig, a research associate at the INP. “The experimental observation of this structure is an important step toward better understanding and controlling plasma jets and their mechanisms of action.”

Image processing at the limits of physics
The demands placed on image processing are extraordinary. A plasma discharge is a highly dynamic phenomenon that changes over time scales of just a few microseconds and extends over a spatial range of only about ten millimeters. Extremely short exposure times are required to make individual discharge channels visible. In this application, exposure times ranging from 9.35 to 30.03 microseconds are used. Images are captured in monochrome as 8-bit single frames. “The key point here is that all cameras must operate in perfect synchronization, as this is the only way to capture the same features within a very short time frame,” emphasizes Artur Wittig. Although two-dimensional single-frame images provide high-resolution views of the discharge, they allow only limited conclusions to be drawn about its spatial structure. In particular, when it comes to self-luminous, highly dynamic objects such as plasma filaments, the actual three-dimensional distribution remains speculative without additional views. Only by capturing images simultaneously from multiple angles is it possible to reliably reconstruct spatial features such as curvature, coiling, or lateral deflection of the discharge.

“We need to ensure that the same plasma filaments are actually captured in every image,” explains Dr. Torsten Gerling, head of the research group. “This requires very precise timing and a high degree of repeatability in relation to the plasma source.”

Calibration of the multi-camera setup on the kINPen using a 3D-printed reference object.

Stable imaging despite highly dynamic discharge
Although a single image taken without a surface may show multiple filaments – so-called guided streamers, which are short-lived, thread-like discharge channels in the plasma – images taken with a surface usually show a clearly dominant discharge path. This behavior is attributed to what is known as the derivational mode: A guided streamer forms a conductive channel to the surface. A kind of transient glow discharge then flashes erratically along this channel. Due to the memory effect, metastable particles from previous discharges facilitate the re-ignition of additional guided streamers. These largely follow the same path, slightly offset by the gas flow.

In particular, when the kINPen is excited at high frequencies, this effect causes the visible plasma structure to form in a spatially reproducible manner across multiple discharges. As a result, it can be reliably captured using imaging techniques.

This physical property provides a fundamental basis for the systematic investigation of highly dynamic plasma discharges using image-processing-based measurement methods.

3D reconstruction of the discharge structure as a point cloud with a centerline (red) to illustrate the discharge channel; normals for orientation (blue).

Multi-view stereo for 3D reconstruction
To experimentally characterize the spatial structure of the plasma discharge, the INP employs a multi-view stereo approach using five IDS cameras operating in sync. The plasma discharge is captured simultaneously from different angles. In addition to precise calibration of the camera system, a reliable spatial reconstruction also requires imaging of the fine discharge structures with as little distortion as possible.

High‑aperture 75 mm lenses from IDS are used, featuring a large 1.2‑inch image circle and an f/2.8 aperture. This optical performance is necessary because the discharge has an axial length of less than 10 mm and a width of less than 1 mm.

“At a viewing distance of about 500 mm, the plasma filament is barely visible; its brightness is roughly equivalent to that of a firefly,” explains Dr. Philipp Mattern, supervisor and reviewer of the master’s thesis conducted at the INP. “Only the combination of sensor and optics makes it possible to achieve high‑quality images despite microsecond exposure times.”

During image analysis, distinctive structures in the plasma discharge are identified and used as cross-image point correspondences, from which the three-dimensional structure of the discharge is reconstructed as a point cloud.

Structure of the self-luminous filament at a distance of 3 mm (exposure time: 40.76 µs).

“The point clouds obtained in this way provide, for the first time, a reliable basis for studying the discharge paths,” explains Artur Wittig. “This allows us not only to visualize the plasma structure, but also to analyze it systematically.”

Camera selection with a focus on triggering and synchronization
The image processing task is handled by five industrial cameras of the uEye CP U3‑31J0CP Rev.  2.2 type from IDS, which are well suited for parallel operation in multi‑camera setups thanks to their triggering and synchronization capabilities.

The conceptual foundation for this setup, as well as the decision to use IDS hardware, was developed by Dr. Philipp Mattern. Some of the scientific and technical support was provided through his engineering firm, M.E.S.S. (Mattern Engineering & Software Solutions). “Based on my experience with similar applications, it was clear that this camera system would be able to meet the demanding optical and temporal requirements,” explains Mattern.

Multi-camera setup with five IDS industrial cameras positioned at 90° angles around the discharge point

The main factors influencing the selection were the capabilities for precise hardware triggering, exact synchronization, and reproducible control of very short exposure times. Due to the highly dynamic nature of the plasma discharge, precise triggering, exact synchronization, and reproducible exposure times in the microsecond range are crucial for capturing the same features in every image. The global shutter sensor used enables distortion-free imaging of the short-lived plasma structure and ensures stable image quality even at exposure times in the microsecond range.
The camera is equipped with a square Sony Pregius S CMOS sensor (IMX546) and offers a resolution of 8.13 megapixels. The combination of a global shutter and backside illumination (BSI) enables short exposure times even under low-light conditions – a key requirement for reliably imaging self-luminous, short-lived plasma structures.

“The comprehensive documentation provided by IDS was also helpful, as was their technical support in designing and validating the configuration of multiple cameras for simultaneous image capture and the stable setup of the multi-camera system,” says Artur Wittig.

Integration is handled via the IDS peak SDK, which enables the configuration and simultaneous operation of multiple cameras. The ability to reliably save and reuse camera settings ensures that experimental measurement series can be conducted under consistent conditions and compared with one another. The control and automation of the multi‑camera setup are handled via the IDS peak API for Python, which conveniently enables parallel operation, triggering, and image storage.

More than just visualization: an experimental proof of concept
The multi-camera methodology developed is not merely intended for illustrative purposes. Rather, it represents an experimental proof of concept: For the first time, it has been demonstrated that the highly dynamic plasma discharge of a kINPen jet can be reconstructed as a three-dimensional point cloud and subsequently analyzed structurally. This provides a practical basis for further investigations into the spatial propagation of plasma jet discharges.

Structural analysis of the plasma discharge in a kINP jet

Furthermore, the method is not limited to kINPen, but can also be applied to other small discharge structures with relatively little effort.

Outlook
The current work continues to focus on the analysis of plasma jet discharges, including under altered operating parameters such as gas flow or discharge mode. Furthermore, other applications are conceivable, particularly in situations where dynamic structures need to be studied with high temporal and spatial resolution. Imaging techniques such as Schlieren or BOS (Background Oriented Schlieren) methods are also the subject of ongoing research. These are optical imaging techniques that do not capture the objects themselves, but rather changes in the fluid, such as air or the working gas. In future, they will open up new possibilities for visualizing invisible flows and density differences in the vicinity of the plasma discharge, thereby complementing the experimental analysis.

IDS’s perspective and technical assessment
The project impressively demonstrates how IDS’s flexible and high-performance image processing solutions are opening up new avenues in experimental research – and thus enabling us to see what was previously invisible. “In applications involving highly dynamic objects such as plasma discharges, it is not individual features that are decisive, but rather the combination of a global shutter sensor and precise, reproducible exposure control via hardware triggering to synchronize multiple cameras,” explains Heiko Seitz, Product Marketing Manager at IDS. “These features make it possible to capture consistent image data even in multi-camera setups, thereby providing a reliable foundation for demanding image processing tasks in research and development.”

Image rights: Leibniz Institut für Plasmaforschung und Technologie e.V. (INP)
© 2026 IDS Imaging Development Systems GmbH

Camera

Model used: U3-31J0CP Rev.2.2
Camera family: uEye CP
Lens used: IDS-20M12-C7528