Dot-Tracking Methodology for Background Oriented Schlieren (BOS)

TL;DR

This paper introduces a dot-tracking methodology for BOS images that enhances accuracy, resolution, and robustness, especially in noisy conditions, by leveraging prior dot pattern information and a correlation correction technique.

Contribution

The paper presents a novel dot-tracking approach for BOS that improves measurement accuracy and robustness, incorporating prior pattern info and a correlation correction for noisy images.

Findings

Achieves near 100% dot detection yield at high densities

Improves displacement estimation accuracy in noisy images

Demonstrates effectiveness with synthetic and experimental BOS data

Abstract

We propose a dot-tracking methodology for processing Background Oriented Schlieren (BOS) images. The method significantly improves the accuracy, precision and spatial resolution compared to conventional cross-correlation algorithms. Our methodology utilizes the prior information about the dot pattern such as the location, size and number of dots to provide near 100% yield even for high dot densities (20 dots/32x32 pix.) and is robust to image noise. We also propose an improvement to the displacement estimation step in the tracking process, especially for noisy images, using a "correlation correction", whereby we combine the spatial resolution benefit of the tracking method and the smoothing property of the correlation method to increase the dynamic range of the overall measurement process. We evaluate the performance of the method with synthetic BOS images of buoyancy driven turbulence rendered using ray tracing simulations, and experimental images of flow in th exit plane of a converging-diverging nozzle.