Imaging techniques for fluid mechanics measurements

Over the centuries, mankind has observed flows through natural phenomena (clouds, smoke, algae...) and created various representations to try and understand their phenomenology. We all remember certain drawings and iconographies by Leonardo da Vinci or Giovanni Battista Venturi, which depict areas of turbulence, or the release of vortices using hand-drawn streamlines. These representations, initially highly qualitative, gradually gave way, over time and with technological progress, to more quantitative characterizations: velocity measurements, trajectographies, tracking of objects in flows, fluid-structure interactions, concentration and temperature measurements, etc., supported by imaging and visualization.

Today, flow characterization in fluid mechanics calls on a wide range of measurement techniques, enabling the exploration of numerous spatial and temporal scales. Among these, imaging-based techniques are undoubtedly among the most fundamental, and they have naturally enjoyed a boom due to their non-intrusive nature and their ability to observe flows at different scales.

Imaging, initially analog, gave way to digital imaging and computer-assisted image processing in the 1980s, bringing this technique into the realm of metrological techniques. The parallel emergence of lasers opened the door to the coupling of techniques and the non-intrusive measurement of increasingly complex flows, particularly multiphase flows. Today, the emergence of artificial intelligence methods, which are revolutionizing imaging and image processing, augurs an even richer future for this technique.

This article presents the state-of-the-art in imaging techniques for fluid mechanics measurement. It begins by describing the different types of experimental set-ups based on imaging, with a detailed description of the equipment encountered and a few reference examples. Next, the issue of signal processing leading to the extraction of quantitative physical information in flows is discussed. Finally, examples drawn from a number of recent research projects are presented for illustrative purposes.