Automatic extraction of wall streamlines from oil-flow visualizations using a convolutional neural network

TL;DR

This paper introduces a convolutional neural network-based method to automatically extract wall streamlines from oil-flow visualizations, significantly reducing analysis time and increasing objectivity.

Contribution

It presents a novel supervised learning approach trained on a large dataset to accurately predict flow directions from oil-flow images, demonstrating high accuracy and generalizability.

Findings

Mean prediction error of 3 degrees on test data

Robust performance across diverse flow configurations

Fast and automated streamline extraction

Abstract

Oil-flow visualizations represent a simple means to reveal time-averaged wall streamline patterns. Yet, the evaluation of such images can be a time-consuming process and is subjective to human perception. In this study, we present a fast and robust method to obtain quantitative insight based on qualitative oil-flow visualizations. Using a convolutional neural network, the local flow direction is predicted based on the oil-flow texture. This was achieved with supervised training based on an extensive dataset involving approximately one million image patches that cover variations of the flow direction, the wall shear-stress magnitude and the oil-flow mixture. For a test dataset that is distinct from the training data, the mean prediction error of the flow direction is as low as three degrees. A reliable performance is also noted when the model is applied to oil-flow visualizations from the literature, demonstrating the generalizability required for an application in diverse flow configurations.