Development of 3D Background Oriented Schlieren with a Plenoptic Camera
Type of DegreeDissertation
MetadataShow full item record
This dissertation investigates a novel idea of implementing background oriented schlieren (BOS) with a plenoptic camera along with its application for 3-dimensional (3D) density measurement and for depth perception (DP) of inhomogeneous structures in an inhomogeneous flow field. In the process, the mathematical implementation of BOS was reworked to address the sensitivity of the measurement technique when used in constrained laboratory conditions. The current equation describing the BOS analysis suggests that the sensitivity of the BOS measurement technique is a function of the focal length of the lens used on the camera. When the conditions of a laboratory were imposed on the BOS experiment, it was found that the f-number on the lens of the camera and the circle of confusion used in setting up the experiment influenced the sensitivity of the experiment. This dissertation also provides guidelines for setting up an effective BOS experiment in a laboratory. Plenoptic cameras are light field based cameras that capture light fields and the images are called plenoptic images. These images once captured can be computationally processed after the fact to generate multiple perspective views. For the first time, in this dissertation, the idea of implementing BOS with the plenoptic camera was explored. The major advantages of BOS with a plenoptic camera is that, multiple perspective are captured with a single camera and the ease with which they are captured. The multiple perspective views obtained from BOS with a plenoptic camera show disparity in the captured images which results in the disparity of the BOS data generated from these images. The ability to generate multiple perspective views from a single plenoptic image and the disparity between them were used in the development of the DP and 3D density measurement capabilities of the BOS with a plenoptic camera. The DP capability of the BOS with a plenoptic camera was successfully demonstrated by measuring the location of a cone shock with less than 3\% error. The cone shock was generated with a cone placed in a Mach 2.0 flow. The 3D density measurement capability was demonstrated with both simulated and physical experiments. From these experiments, it was learned that the BOS with a plenoptic camera can accurately measure the 3D density distribution of an inhomogeneous flow field as long as the angle between the inhomogeneous structure and the optical axis of the camera is in the range of the viewing angles of the camera. But, as all the perspective views generated from the plenoptic image are captured through the aperture of the camera lens, the range of the viewing angles obtained in the perspectives is small. This small range of viewing angles limits the 3D density measurement applicability of the BOS with a plenoptic camera.