AcoustRivNN proposes to develop a system to estimate the flow and the granulometry of the sediment transport in a river from the acoustic pressure generated by the latter using methods from artificial intelligence.

The estimation of the sediment flow carried by water, in rivers or estuaries, is a crucial issue for the management of the latter, allowing to carry out scientific studies, restoration or prevention projects, as well as operational works.

Given the lack of effective methods to estimate the flow of sediment, the AcoustRivNN project proposes to provide a "proof of concept" by developing an original system based on deep learning to estimate the flow of coarse sediments from the simple acoustic pressure generated by the latter and measured by hydrophones.

The originality of this project lies, in particular, in its interdisciplinary aspect proposing to adapt methods from artificial intelligence particularly effective in many applied fields. This project, which is part of the transversal axis 2: "Observations/information systems/modeling" of the ECCOREV federation, is structured in two phases:

- Phase I proposes to build an acoustic database referenced in the laboratory necessary for the training of a neural network.

- Phase II aims to develop a neural network model to characterize the acoustics of sediment flow.

DOI: https://doi.org/10.34930/dc3225de-ef03-4134-927e-2347d75d8b41

Citation:

Gassier, G., Michal, T., & Dussouillez, P. (2022). AcoustRivNN : flow and the granulometry of the sediment transport in a river from acoustic pressure [Data set]. CEREGE UMR 7330 CNRS. https://doi.org/10.34930/DC3225DE-EF03-4134-927E-2347D75D8B41