Underwater Optical-Sonar Image Fusion Systems

Area of Science:

• Marine robotics and imaging technology.
• Computer vision and sensor fusion.

Background:

• Unmanned underwater operations rely on optical and sonar sensors.
• Underwater images suffer from noise and distortion, hindering intuitive understanding.
• Existing imaging methods lack integrated color and distance information.

Purpose of the Study:

• To develop an optical and sonar image fusion system for unmanned underwater vehicles.
• To integrate color information from optical images and distance information from sonar images.
• To enhance the visualization and understanding of underwater environments.

Main Methods:

• Developed a system for fusing optical and sonar images using calibrated transformation matrices.
• Applied median filter and gamma correction to improve contrast and reduce noise.
• Utilized underwater image quality measure (UIQM) and underwater color image quality evaluation (UCIQE) for performance evaluation.

Main Results:

• Image fusion significantly improved image quality, increasing mean UIQM by 94% and UCIQE by 27%.
• Contrast-to-noise ratio improved sixfold after image pre-processing.
• Achieved good spatial agreement with an average Intersection over Union (IoU) of 75% between optical and sonar pixels.

Conclusions:

• The proposed optical-sonar fusion system effectively integrates color and distance data.
• Enhanced image quality and spatial agreement improve underwater situation awareness for unmanned operations.
• This technology aids in better visualization for complex underwater tasks.