Model-based adaptive 3D sonar reconstruction in reverberating environments.

Area of Science:

• Geophysics
• Oceanography
• Signal Processing

Background:

• Underwater environments, especially shallow waters, present significant challenges for sonar-based bathymetry due to multipath echoes and volume reverberation.
• These acoustic artifacts lead to inaccurate depth estimations, hindering effective 3D underwater scene reconstruction.

Purpose of the Study:

• To develop a novel model-based approach for enhancing 3D bathymetry reconstruction using side scan sonar arrays.
• To address the limitations of existing methods in complex and reverberating underwater environments.

Main Methods:

• Proposed an adaptive filter utilizing multiple original geometrical models for processing side scan sonar data.
• Implemented a multimodel approach for tracking and separating direction of arrival trajectories of multiple echoes.
• Incorporated prior information on echo temporal evolution for a model-based echo tracking stage to reject clutter.

Main Results:

• The proposed filter demonstrated clutter-free and regularized bathymetric reconstruction on both simulated and real sonar data.
• Successfully tracked and separated multiple echo trajectories, significantly improving depth estimation accuracy.
• Model validation using goodness of fit tests confirmed the effectiveness of the model-based processing.

Conclusions:

• The developed model-based adaptive filter significantly improves 3D bathymetry reconstruction in challenging underwater environments.
• The multimodel approach is crucial for accurately separating echoes and mitigating reverberation effects.
• This method offers a robust solution for accurate underwater mapping using side scan sonar.