TY - JOUR
T1 - Seabed classification using physics-based modeling and machine learning
AU - Frederick, Christina
AU - Villar, Soledad
AU - Michalopoulou, Zoi Heleni
N1 - Funding Information:
The authors are grateful to Kyunghyun Cho, Center for Data Science, New York University Division of Mathematical Sciences (CDS, NYU) for valuable discussions. The research of C.F. is supported by the National Science Foundation (NSF) DMS Grant No. 1720306. The research of Z.H.M. is supported by the Office of Naval Research (ONR) Grant Nos. N000142012029 and N000141812125. S.V. is partly supported by NSF DMS Grant Nos. 1913134, EOARD FA9550-18-1-7007, and the Simons Algorithms and Geometry (A and G) Think Tank. The project started while S.V. was with the NYU Center for Data Science and Courant Institute of Mathematical Sciences.
Publisher Copyright:
© 2020 Acoustical Society of America.
PY - 2020/8/1
Y1 - 2020/8/1
N2 - In this work, model-based methods are employed, along with machine learning techniques, to classify sediments in oceanic environments based on the geoacoustic properties of a two-layer seabed. Two different scenarios are investigated. First, a simple low-frequency case is set up, in which the acoustic field is modeled with normal modes. Four different hypotheses are made for seafloor sediment possibilities, and these are explored using both various machine learning techniques and a simple matched-field approach. For most noise levels, the latter has an inferior performance to the machine learning methods. Second, the high-frequency model of the scattering from a rough, two-layer seafloor is considered. Again, four different sediment possibilities are classified with machine learning. For higher accuracy, one-dimensional convolutional neural networks are employed. In both cases, the machine learning methods, both in simple and more complex formulations, lead to effective sediment characterization. The results assess the robustness to noise and model misspecification of different classifiers.
AB - In this work, model-based methods are employed, along with machine learning techniques, to classify sediments in oceanic environments based on the geoacoustic properties of a two-layer seabed. Two different scenarios are investigated. First, a simple low-frequency case is set up, in which the acoustic field is modeled with normal modes. Four different hypotheses are made for seafloor sediment possibilities, and these are explored using both various machine learning techniques and a simple matched-field approach. For most noise levels, the latter has an inferior performance to the machine learning methods. Second, the high-frequency model of the scattering from a rough, two-layer seafloor is considered. Again, four different sediment possibilities are classified with machine learning. For higher accuracy, one-dimensional convolutional neural networks are employed. In both cases, the machine learning methods, both in simple and more complex formulations, lead to effective sediment characterization. The results assess the robustness to noise and model misspecification of different classifiers.
UR - http://www.scopus.com/inward/record.url?scp=85090181396&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85090181396&partnerID=8YFLogxK
U2 - 10.1121/10.0001728
DO - 10.1121/10.0001728
M3 - Article
C2 - 32873029
AN - SCOPUS:85090181396
SN - 0001-4966
VL - 148
SP - 859
EP - 872
JO - Journal of the Acoustical Society of America
JF - Journal of the Acoustical Society of America
IS - 2
ER -