2023_programme: Multi-spectral multibeam imaging of palaeo-ice stream landforms in the Baltic Sea

Back


  • Session: 06. Marine sediment acoustics
    Organiser(s): Megan Ballard, Kevin Lee and Nick Chotiros
  • Lecture: Multi-spectral multibeam imaging of palaeo-ice stream landforms in the Baltic Sea
    Paper ID: 2018
    Author(s): Tegowski Jaroslaw, Trzcinska Karolina, Kasula Marcelina, Paszkuta Marcin
    Presenter: Tegowski Jaroslav
    Abstract: This research presents the preliminary results of the Baltic STREAMBAL project uses imaging and mapping of the seafloor with multibeam echosounder to identify palaeo-ice streams activity and their geomorphic traces in the Central and Southern Baltic Sea. Recognition and classification of bottom morphological forms was carried out through parameterization of the bathymetric model and maps of the relative backscattering intensity. The Digital Elevation Model was subjected of two-dimensional Fourier Analysis allowing calculation of spectral parameters such as spectral width, spectral skewness and spectral moments. In addition, an autocorrelation radius was calculated using the Wiener-Khinchin theorem to quantify bottom roughness and assess the regularity of glacial relief. A novel algorithm was also developed to compensate for the angular dependence of echo intensity on the bottom in order to generate a uniform backscatter maps. The calculated parameters were the input set for the object-based image analysis (OBIA) algorithm enabling semi-automatic separation of palaeo-ice stream landforms on the use of the theory of moments with invariants. The values of moments with invariants, determined on the basis of the available maps, create a new, valuable set of data, which are the geometrical parameters of the scene representing the landforms. The obtained data sets are used in machine learning methods, deep machine learning methods, etc. The effectiveness of the method is checked on the basis of comparing the entropy results of the input maps. This work was supported by the National Science Centre (Poland), project no 2021/41/B/ST10/01086.
      Download the full paper
  • Corresponding author: Prof Jaroslaw Tegowski
    Affiliation: University of Gdańsk
    Country: Poland
    e-mail:

Back