Multibeam Seabed Classification Project

Background
For half a century hydrographers have used single beam echosounders to chart water depths. Over the past decade multibeam technology has evolved into the tool of choice for bathymetric mapping. Providing large area coverage, multibeam sonars cover swaths more than four times water depth. Significant efficiencies in vessel time requirements are gained.

Objective
The bathymetric data provided by a multibeam system does not directly indicate composition of the substrate surveyed. But, as with single beam sounders, there is more information in the multibeam echoes than simply water depth. Backscatter data, influenced by the reflectivity of the seabed can be extracted from the signal and used for seabed classification.

Quester Tangent currently uses a unique suite of algorithms and a data management system in the commercially available QTC VIEW ™ and QTC IMPACT ™ products for single beam applications. The purpose of the MBSC project is to further develop this technology to classify the signals from multibeam sonars.

The MBSC project has developed, implemented and validated a remote seabed classification system that will be applicable to most multibeam sonars, providing seabed classification over the multibeam swath. This will add significant value to the multibeam sonar, at incremental cost. Survey information will be greatly enhanced for hydrographic, fisheries habitat, military, offshore site survey and dredging purposes, and will be invaluable to all current and future users of multibeam systems.

QTC MULTIVIEW ™ Version 3.00 is the current release and it provides image-based classification of multibeam data. QTC SIDEVIEW, which classifies sidescan sonar data, is also available. Customers in Canada, Ireland, the United States and Korea are using QTC MULTIVIEW for classification of Simrad and Reson multibeam data.

Partners
To achieve the project goals, Quester Tangent has built upon existing cooperative research agreements and strategic partnerships. The project is supported by partnerships with:

• Canadian Centre for Marine Communications (CCMC)
• Canadian Department of National Defence Industrial Research Program (DIR)
• The Institute for Pacific Ocean Science and Technology (IPOST)
• Western Economic Diversification Fund (WED)

Single beam Classification
Methods of characterising the sea floor include:

• Cameras
• Physical Sampling
• Side-scan sonars
• Sub-bottom sonars

These methods are generally slow, require specialised sample analysis and data interpretation personnel. They are relatively expensive in terms of service and operating costs.

Remote sensing via acoustic seabed classification significantly reduces these costs. Remote seabed classification requires an acoustic data acquisition system, a set of algorithms to analyse the data, determine acoustic class, and relate the classification results to the physical characteristics of the sediments. Classification requires an ordination technique to group echoes with similar features.

Visit our QTC VIEW ™ and QTC IMPACT ™ pages for more information on single beam technology and products.

Multibeam Classification

Multibeam sonar technology has evolved rapidly. The systems now provide complete survey coverage of the sea floor with marginally higher operating costs than more conventional methods. Capital costs have traditionally been higher, and the post-processing and service costs are significant.

The multibeam project will take single beam processing, modify existing algorithms and data handling techniques, build new algorithms as necessary, and produce a multibeam seabed classification product. This is now possible because recent developments in electronics, computers and manufacturing techniques have lowered equipment costs and reduced post-processing times to acceptable levels for some applications.

Project Goals
The project goals in summary:

• Evaluate existing seabed classification algorithms and create new signal processing techniques to extract seabed information from multibeam sonar waveforms.
• Multibeam sonar signals are synthetic waveforms produced by beam forming techniques. Existing echo classification algorithms will be applied to these signals to evaluate their effectiveness. Limitations will be identified and new algorithms created to overcome the deficiencies. Constraints to classification from the multibeam sonar data generation will be identified and removed wherever possible.
• Develop new algorithms for multibeam image data conditioning and feature extraction.
• Redesign Quester Tangent's existing acoustic data acquisition system to capture, condition and handle data received from multibeam sonar.
• There is a fundamental difference in the nature of a sonar waveform as sound is projected at shallower angles. Conversion of the existing acquisition hardware is necessary for preservation of signal shape characteristics at shallow viewing angles.
• Develop a data management system to facilitate processing of multibeam sonar data for seabed classification. Classification of multibeam data requires the integration of seabed information separated in both space and time. Processing the seabed waveform to extract characteristics related to the physical properties of the seabed will require multi-channel data conduits with enhanced communications to handle large data volumes.
• Validation of the classification system by conducting experimental surveys in areas with extensive ground-truth and within operational water depths mandated by market sector requirements.