Related Experiment Video
Updated: Mar 24, 2026
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
858
A Robust and Fast Method for Sidescan Sonar Image Segmentation Using Nonlocal Despeckling and Active Contour Model
IEEE Transactions on Cybernetics
|March 16, 2016
Summary
This study introduces a new method for segmenting sidescan sonar images, effectively reducing speckle noise and intensity variations for better underwater object detection. The approach ensures accurate and efficient image analysis.
Area of Science:
- Marine technology
- Image processing
- Underwater acoustics
Background:
- Sidescan sonar image segmentation is crucial for underwater object detection and recognition.
- Speckle noise and intensity inhomogeneity present significant challenges in sonar image analysis.
Purpose of the Study:
- To propose a robust, fast, and accurate method for sidescan sonar image segmentation.
- To address the limitations of existing methods in handling noise and intensity variations.
Main Methods:
- Integration of nonlocal means-based speckle filtering (NLMSF) for noise reduction.
- Coarse segmentation using k-means clustering to reduce iterations.
- Fine segmentation with an improved region-scalable fitting (RSF) model incorporating edge-driven constraints.
Main Results:
- The proposed method effectively removes speckle noise while preserving image details.
- It successfully segments noisy and inhomogeneous sonar images.
- Experimental results show robustness against noise and intensity variations, with high speed and accuracy.
Conclusions:
- The developed method offers a significant improvement for sidescan sonar image segmentation.
- It provides a reliable tool for underwater object detection and recognition tasks.
- The integration of NLMSF, k-means, and improved RSF model enhances segmentation performance.
