Underwater sonar image detection: A combination of non-local spatial information and quantum-inspired shuffled frog...

Area of Science:

• Signal Processing
• Artificial Intelligence
• Computer Vision

Background:

• Sonar image analysis presents challenges in object detection due to noise and parameter sensitivity.
• Existing non-local spatial information methods often suffer from inappropriate filtering parameters, impacting denoising performance.
• Efficient and accurate underwater object detection is crucial for various applications.

Purpose of the Study:

• To propose a novel method for underwater object detection in sonar images.
• To address the issue of inappropriate filtering degree parameters in non-local spatial information processing.
• To develop an efficient and accurate detection algorithm using a quantum-inspired shuffled frog leaping algorithm with a new search mechanism.

Main Methods:

• A novel filtering degree parameter was developed to optimize non-local spatial information filtering for sonar images.
• A quantum-inspired shuffled frog leaping algorithm based on a new search mechanism (QSFLA-NSM) was proposed.
• Real number encoding for frog individuals, a combined fitness function (intra-class and inter-class differences), and an improved search mechanism inspired by QPSO and SFLA were utilized.

Main Results:

• The proposed QSFLA-NSM demonstrated precise and rapid detection of sonar images.
• The method effectively solved the problem of inappropriate filtering degree parameters.
• Comparative experiments confirmed the effectiveness and adaptability of the proposed method on sonar images, UCI datasets, and benchmark functions.
• Reduced time complexity was achieved.

Conclusions:

• The combined approach of optimized non-local spatial information and QSFLA-NSM offers a significant advancement in underwater object detection.
• The developed method provides improved accuracy, speed, and robustness compared to existing techniques.
• The study highlights the potential of quantum-inspired algorithms in complex image analysis tasks.