Quantum bathymetry

Bathymetry is the study of water depths, conducted by S[&]T using remote sensing data. This is important for the Dutch Ministry of Defence to gather useful intelligence. They use it, for example, to determine possible landing sites for amphibious vehicles or the presence of sandbanks. The depth of the ocean floor between 0 and 20 metres is derived from large amounts of multispectral satellite data using advanced machine learning algorithms. These data, consisting of 13 spectral bands with a temporal resolution of 5 days and a spatial resolution between 10 and 60 metres, result in an enormous amount of data that becomes unmanageable for classical methods.

Quantum neural networks can achieve high accuracies with small amounts of data because quantum machine learning algorithms can generalise well with little data. The challenge is loading large datasets onto a quantum processor. To make optimal use of quantum hardware, classical dimensionality reduction was applied to reduce the high spectral dimensionality of the available data. The data were then embedded in a variational quantum classifier to classify them into depth intervals.

The quantum algorithm was executed on a 4-qubit simulator and tested on real multispectral data near the Mediterranean Sea. The classical results were reproduced with a quantum classifier. Measuring water depth more accurately with less data is beneficial for climate monitoring, resource exploration, and navigation.