Computational challenges in processing and analysis of full-watercolumn multibeam sonar data.

Several multibeam sonar systems are now capable of collecting and recording data samples covering the full-watercolumn, not just the seabed. Such systems, while still facing hardware challenges such as limited dynamic range and bandwidth, collect vast quantities of data, generally an order of magnitude more than conventional hydrographic multibeam or scientific single beam sonar systems. In this paper, the challenges faced by data processing systems for analysis of full-watercolumn multibeam sonar data are explored. Full-watercolumn multibeam data sets are valuable to scientists from traditionally diverse fields, providing simultaneous information about bathymetry, seabed type and habitats, and biomass in the watercolumn. Aspects of the data processing pipeline that are considered in this paper include raw data storage, data pre-processing, visualization and exploratory data analysis, statistical data analysis and post-processing, and presentation and interpretation of results. A general framework is outlined, and specific aspects applicable to the kind of data and problems at hand are emphasized. Proposed solutions to some of the challenges are reviewed and placed within an overall framework of multibeam sonar watercolumn data analysis. It will become clear that successful contributions to the field have been made, but that a general analysis method has yet to emerge.