Improvements to the Automatic Calibration of Digital Focusing Color Schlieren Systems

Abstract

As a diagnostic method, Schlieren imaging offers the unique ability to visualize flows with a completely non-intrusive method. By visualizing small changes in the refractive index of a fluid, Schlieren imaging is able to visualize changes in temperature, pressure, and composition. While historic Schlieren methods have been limited to providing largely qualitative information, modern Schlieren methods are able to extract quantitative information from flow phenomena. This is possible through either Background Oriented Schlieren (BOS) or Digital Focusing Color Schlieren (DFCS) methods. While BOS is relatively flexible, it requires a large amount of computational effort to process each frame of captured information. While DFCS does not require per-frame computation, DFCS systems do require a precise calibration of the location, size, and rotation of a digital background and cutoff grid relative to an imaging sensor. Existing calibration algorithms use pixel-space techniques to perform this calibration. These algorithms are prone to error, sensitive to noise, and often unable to provide the required level of accuracy without manual intervention. Several improvements to the calibration of a DFCS system have been implemented which overall produce a significant improvement to the accuracy and resilience of DFCS calibration. The first of these improvements are new calibration methods utilizing both frequency and pixel space methods capable of calibrating the cutoff grid and background display with a significantly higher level of accuracy. Secondly, these new methods have been extended to work with hexagonal grids which are capable of providing a higher sensitivity and superior color mixing properties. Finally, a new method of producing backgrounds transformed for use on displays with non-square pixels has been implemented, and testing has shown that this is required to properly use some displays. This improved and extended DFCS system provides a level of sensitivity higher than previously possible.