Plenoptic BOS: Combining a Plenoptic Camera with the Background Oriented Schlieren Technique

Abstract

The background oriented schlieren (BOS) technique is a method of observing the refraction of light rays passing through a density varying medium. When used with a conventional imaging system, this technique captures a single line-of-sight integrated quantity that is a result of the refractive index gradients associated with the density varying medium. With the inherent three-dimensionality of most flow fields, an alternative imaging system called the plenoptic camera has been introduced to the scientific community as a viable option in 3D BOS systems. This camera has the ability to capture 3D information in a single snapshot, which allows for images of different perspectives and different focal planes to be generated from a single raw plenoptic image. The integration of a plenoptic camera into the BOS technique, termed plenoptic BOS, is discussed in this work. The use of a plenoptic camera in a BOS setup provides the ability to acquire multiple line-of-sight integrated quantities, which can be used to create a BOS light field similar to that of the original light field captured by the plenoptic camera. This type of structured BOS light field allows for the ability to computationally generate focused schlieren images through the volume. The experiments used to explore this new technique include an arrangement with two flames placed at different depths and a setup of a single heated jet placed at ten different locations within the field of view. The former arrangement provided the ability to qualitatively highlight gradients being produced at different depths in 3D space. The latter arrangement explored the sensitivity of the BOS setup with respect to the jet's position relative to the background as well as the ability to quantify the jet's position in 3D space based on the collected measurements. Such results provide motivation to further explore the plenoptic BOS technique, improve the algorithms associated with depth estimation, and ultimately progress towards a truly 3D reconstruction of a density field.