Related Experiment Video
Updated: Oct 13, 2025
09:19
Measuring the Structure, Composition, and Change of Underwater Environments with Large-area Imaging
Published on: April 18, 2025
902
A Matching Algorithm for Underwater Acoustic and Optical Images Based on Image Attribute Transfer and Local Features
Xiaoteng Zhou1, Changli Yu1, Xin Yuan1
1School of Ocean Engineering, Harbin Institute of Technology, Weihai 264209, China.
Sensors (Basel, Switzerland)
|November 13, 2021
Summary
This study introduces a novel method for matching sonar and optical images, crucial for underwater vision. The technique effectively preprocesses multimodal images, enabling accurate matching for enhanced multisensor information fusion.
Area of Science:
- Robotics and Computer Vision
- Underwater Imaging Systems
- Sensor Fusion Technologies
Background:
- Underwater vision systems commonly use sonar and optical cameras, but matching images from these sensors is challenging due to differing modalities and local features.
- Existing general matching methods fail with sonar and optical images because their independent imaging mechanisms create significant feature discrepancies.
- Effective multisensor information fusion (MSIF) relies on accurate image matching between diverse sensor inputs.
Purpose of the Study:
- To address the challenge of acousto-optic image matching in underwater environments.
- To develop a robust method for aligning sonar and optical images for improved MSIF.
- To leverage image attribute transfer and advanced local feature descriptors for multimodal image matching.
Main Methods:
- An image attribute transfer algorithm was employed to bridge the modality gap between sonar and optical images.
- Advanced local feature descriptors were utilized to extract and match salient features across different image types.
- The proposed method was tested using both real and simulated underwater image datasets.
Main Results:
- The developed approach effectively preprocesses multimodal underwater images, overcoming inherent feature differences.
- Accurate matching results were achieved between sonar and optical images, validating the method's efficacy.
- The technique demonstrated superior performance compared to general matching methods in multimodal scenarios.
Conclusions:
- The proposed image attribute transfer and feature descriptor method offers a viable solution for underwater acousto-optic image matching.
- This advancement facilitates better utilization of data from sonar and optical sensors in underwater applications.
- The work contributes a new approach to the field of multisensor information fusion for underwater exploration and monitoring.
