Unconstrained Synthetic Aperture Sonar
Final Report Abstract
Sonar is an essential sensor underwater as it also works under poor or non-existent visibility conditions and over longer distances. However, its spatial resolution is based on a combination of several transducers, i.e., a scanning beam is (roughly) approximated by interferences. A higher number of transducers over a larger area leads to a better spatial resolution. However, this is limited by, e.g., size, power consumption and cost. A popular approach is therefore to use a synthetic aperture, i.e. the sonar is positioned with its N transducers at k locations, resulting in a virtual sensor with kN transducers. The state of the art for this Synthetic Aperture Sonar (SAS) is strongly coupled to constraints, e.g., using k equidistant positions on a line perpendicular to the orientation of the sonar. In this project, the foundations for an unconstraint SAS were investigated, i.e. a SAS that (a) can be calculated on arbitrary trajectories and (b) does not require navigation data (GPS, INS, etc.). Among others, it was shown that the basic idea of using just registration of the raw data as the basis for SAS is not only possible without navigation data, but that it even leads to an improvement, especially in the near field, with accordingly adapted methods. Furthermore, Synthetic Scan Formation (SSF) was introduced as a new method that works on a level above the signal processing of classic synthetic aperture methods and below Simultaneous Localization and Mapping (SLAM). In particular, Synthetic Scan Formation can be used directly with commercial, off-the-shelf sonars without requiring access to the raw data of the transducers. It is also very efficiently computable and leads to clear improvements over SLAM with sonar according to the state of the art. Also, contributions to robust registration methods as an alternative to standard navigation methods have been made by among others the development of Fourier-SOFT in 2D (FS2D). FS2D uses the decoupling of rotation and translation in the frequency domain known from Fourier-Mellin. The difficult part of determining the rotation parameter is solved in FS2D with a projection of the Fourier magnitude onto a sphere and the SO(3) Fourier transform (SOFT). In addition to validating the project results in realistic simulations, experiments were also carried out in pools with ground truth localization and in field tests, e.g., in the Denkort U-Boot Bunker Valentin, which demonstrated their usefulness in practice. The project results are also of interest for applications outside of marine systems and sonar. For example, a variant of spectral 3D registration with Fourier-Mellin-SOFT (FMS) was developed for magnetic resonance tomography (MRT) data, which can, e.g., be used to detect and localize hand bones.
Publications
-
A Divide and Conquer Method for 3D Registration of Inhomogeneous, Partially Overlapping Scans with Fourier Mellin SOFT (FMS). 2020 IEEE International Conference on Robotics and Automation (ICRA), 8594-8601. IEEE.
Bulow, Heiko; Mueller, Christian A.; Gomez, Chavez Arturo; Buda, Frederike & Birk, Andreas
-
Synthetic Aperture Sonar (SAS) without Navigation: Scan Registration as Basis for Near Field Synthetic Imaging in 2D. Sensors, 20(16), 4440.
Bülow, Heiko & Birk, Andreas
-
Registration of Magnetic Resonance Tomography (MRT) Data with a Low Frequency Adaption of Fourier-Mellin-SOFT (LF-FMS). Sensors, 21(8), 2581.
Bülow, Heiko & Birk, Andreas
-
Seeing through the forest and the trees with drones. Science Robotics, 6(55).
Birk, Andreas
-
Synthetic Scan Formation for Underwater Mapping With Low-Cost Mechanical Scanning Sonars (MSS). IEEE Access, 11, 96854-96864.
Hansen, Tim & Birk, Andreas
-
Underwater Exploration with Sonar of the Flooded Basement of a WW-II Submarine Bunker in the Context of the Digitization of Cultural Heritage. OCEANS 2023 - Limerick, 1-6. IEEE.
Hansen, Tim; Buda, Frederike & Birk, Andreas
-
Using Registration with Fourier-SOFT in 2D (FS2D) for Robust Scan Matching of Sonar Range Data. 2023 IEEE International Conference on Robotics and Automation (ICRA), 3080-3087. IEEE.
Hansen, Tim & Birk, Andreas
-
An Open-Source Solution for Fast and Accurate Underwater Mapping with a Low-Cost Mechanical Scanning Sonar. 2024 IEEE International Conference on Robotics and Automation (ICRA), 9968-9975. IEEE.
Hansen, Tim & Birk, Andreas
-
Scanning Sonar Data From an Underwater Robot With Ground Truth Localization. IEEE Access, 12, 129202-129211.
Hansen, Tim; Belenis, Bogdan; Firvida, Miguel Bande; Creutz, Tom & Birk, Andreas
