2019_programme: HIGH-RESOLUTION LOW-FREQUENCY COMPRESSIVE SAS IMAGING WITH DISTRIBUTED OPTIMIZATION

• Session: 13. Machine learning, compressive sensing and signal processing
  Organiser(s): Xenaki Angeliki, Gerstoft Peter
• Lecture: HIGH-RESOLUTION LOW-FREQUENCY COMPRESSIVE SAS IMAGING WITH DISTRIBUTED OPTIMIZATION [invited]
  Paper ID: 876
  Author(s): Xenaki Angeliki, Pailhas Yan, Hamon Ronan
  Presenter: Xenaki Angeliki
  Presentation type: oral
  Abstract: Synthetic aperture sonar (SAS) provides very high resolution acoustic imaging by processing coherently the backscattered acoustic signal recorded over several consecutive pings. Utilizing wideband high-frequency pulses, synthetic aperture processing can achieve centimetric range and cross-range resolution resulting in optical-like images of the seafloor reflectivity. Low frequency waves can be partly transmitted through objects or penetrate the seafloor providing information about internal structure and buried objects. Such frequency-specific information is crucial for improving the performance of automatic target recognition algorithms. \nBoth sub-band processing and data reduction, however, degrade the SAS resolution. In this paper, SAS imaging is formulated as an l1-norm regularized least-squares optimization problem which improves the resolution by promoting parsimonious representations of the data. \nWe use an algorithm based on the alternating direction method of multipliers (ADMM) to solve the optimization problem in a distributed and computationally efficient way. The resulting SAS image is the consensus outcome of collaborative filtering of the data from each ping. \nThe potential of the proposed method for high-resolution, low-frequency SAS imaging is demonstrated with simulated data.
    Download the full paper
• Corresponding author: Dr Xenaki Angeliki
  Affiliation: STO-CMRE, NATO
  Country: Italy
  e-mail: