Full-field optical visualization techniques in “dilute” particle-laden flows

Abstract An overview is presented of camera-based techniques used in “dilute” (optically transparent) particle-laden flows for the quantitative visualization of both flow field and dispersed particles. Camera-based techniques nowadays available to experimentalists comprise both mature and newly developed techniques, and depending on technical expertise, available budget, and research goals, this review shows the wide variety of techniques to choose from. First, the principles of light scattering from particles are reviewed after which different methods to distinguish between dispersed particles and fluid phase are discussed including fluorescence and refractive index matching techniques. The camera-based techniques are divided into (i) direct imaging techniques and (ii) computational cameras. The first category comprises well-established techniques such as laser-/LED-based particle image velocimetry and shadowgraphy, whereas the second category discusses digital holography and recent newly developed plenoptic cameras and diffuser-based imaging. The latter two represent novel single-camera techniques whose usage in particle-laden flows has yet to be established. Since camera-based particle-laden flow measurements lead to large data sets that are cumbersome and time-consuming to process, an overview of recently developed and applied machine learning techniques is given. These have already made an impact in the processing of digital holography results, and especially physics-informed neural networks are expected to make an impact in particle-laden flow analysis reducing the need for large “ground truth” data sets. Finally, challenges associated with experimental setup and cost, spatial/temporal resolution requirements, segregation between dispersed and fluid phase as well as data processing are discussed.