AUTHOR=Chen Yonghao , Liu Xiaoyun , Jiang Jinyang , Gao Siyu , Liu Ying , Jiang Yueqiu TITLE=Estimation of ocean turbulence intensity using convolutional neural networks JOURNAL=Frontiers in Physics VOLUME=11 YEAR=2023 URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2023.1279476 DOI=10.3389/fphy.2023.1279476 ISSN=2296-424X ABSTRACT=

Understanding the transmission of light in ocean turbulence is of great significance for underwater communication, underwater detection, and other fields. The properties of ocean turbulence can affect the transmission characteristics of light beams, therefore it is essential to estimate the ocean turbulence intensity (OTI). In this study, we propose a deep learning-based method for predicting the OTI. Using phase screens to simulate ocean turbulence, we constructed a database of distorted Gaussian beams generated by Gaussian beams passing through ocean turbulence with varying intensities. We built a convolutional neural network and trained it using this database. For the trained network, inputting a distorted beam can accurately predict the corresponding intensity of ocean turbulence. We also compared our designed network with traditional network models such as AlexNet, VGG16, and Xception, and the results showed that our designed network had higher accuracy.