Abstract
This paper presents the use of light field data, recorded in a snapshot from a single plenoptic camera, for 3-D visualization of transparent fluid flows. We demonstrate the transfer of light field deconvolution, a method so far used only in microscopy, to macroscopic scales with a photographic setup. This technique is suitable for optically thin media without any additional particles or tracers and allows volumetric investigation of non-stationary flows with a simple single camera setup. An experimental technique for the determination of the shift-variant point spread functions is presented, which is a key for applications using a photographic optical system. The paper shows results from different test cases with increasing complexity. Reconstruction of the 3-D positions of randomly distributed light points demonstrates the achievable high accuracy of the technique. Gas flames and droplets of a fluorescent liquid show the feasibility of the proposed method for the visualization of transparent, luminous flows. The visualizations exhibit high quality and resolution in low-contrast flows, where standard plenoptic software based on computer vision fails. Axial resolution depends on the data and is about an order of magnitude lower than the lateral resolution for simple point objects. The technique also allows the time-resolved analysis of flow structures and the generation of 3D3C-velocity fields from a sequence of exposures.Graphical
Highlights
Fluid flows involve complex three-dimensional structures and interactions on different spatial scales, and their volumetric investigation and visualization are of utmost interest for the understanding of flow physics
The article is organized as follows: After a brief introduction to plenoptic cameras and light field deconvolution in Sect. 2, we present our approach for the experimental determination of the point spread function (PSF) in Sect. 3, which is a key for the calibration of the optical system
All volume reconstructions were carried out using the MATLAB code published by Prevedel et al (2014), which is an implementation of the Richardson–Lucy-based deconvolution method proposed by Broxton et al (2013). It has been pointed out by several authors that sampling of the light field by a plenoptic camera is not uniform throughout the volume, but changes significantly with the depth coordinate (Bishop and Favaro 2009; Broxton et al 2013; Stefanoiu et al 2019)
Summary
Fluid flows involve complex three-dimensional structures and interactions on different spatial scales, and their volumetric investigation and visualization are of utmost interest for the understanding of flow physics. Modern techniques are often based on tomography, where the volume is reconstructed from multiple projections under different viewing angles. Exploiting flow parameters such as refractive index, luminosity or the positions of added tracers allows to derive volumetric results from a set of 2D measurements (Cai et al 2013). A plenoptic camera features an additional array of microlenses (MLA), placed at some distance in front of the sensor, which allows to capture this lost information (Lippmann 1908): Depending on its direction, the microlenses distribute incoming light to different locations on the image sensor beneath. Appropriate decoding of the recorded data allows to extract the light field of the object
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
