Machine learning has proven to be remarkably successful in classifying images. In stratified flow, changes in the index of refraction of light due to variation in the local density can be used to "visualize" the density field with a number of optical techniques. Of particular interest is the very large collection of annotated shadowgraphs from the SID experiment. The images cover a number of typical situations encountered in stratified turbulence (from laminar and quasi-laminar flows to fully turbulent flows). This collection of shadowgraphs offers an unprecedented opportunity to consider some fundamental questions ranging from the capacity of an AI model to classify shadowgraphs based on the nature of the flow, to predicting ahead of time when turbulence will occur in a flow, to transferring learning from shadowgraphs to other optical techniques (e.g., BOS or PIV).
This work is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. Any reuse must credit the author(s) and provide a link back to this page.