Object classification and acoustic imaging with active sonar

Area of Science:

• Acoustics
• Signal Processing
• Machine Learning

Background:

• Underwater acoustic classification and imaging are crucial for various applications.
• High-frequency active sonar systems present unique challenges for accurate detection and identification.
• Existing methods often lack a unified theoretical framework for optimal performance.

Purpose of the Study:

• To develop a comprehensive Bayesian theoretic framework for underwater acoustic classification and imaging.
• To present and evaluate optimum decision rules and array processing techniques.
• To establish a systematic methodology for performance evaluation.

Main Methods:

• Bayesian inference and decision theory applied to acoustic signal processing.
• Development of optimal array processing algorithms for sonar data.
• Derivation of a performance evaluation methodology incorporating object geometry and a priori knowledge.

Main Results:

• The study presents a unified Bayesian framework for practical acoustic classification systems.
• Optimum decision rules and array processing strategies are derived and evaluated.
• New quantitative results link classifier performance to object geometry, acoustic imaging, and prior knowledge accuracy.

Conclusions:

• The developed Bayesian framework provides a robust foundation for high-frequency active sonar classification and imaging.
• The derived methodology enables systematic performance evaluation and optimization.
• Understanding the influence of object geometry and prior knowledge is key to improving underwater acoustic systems.