Real-time super-resolved 3D in turbid water using a fast range-gated CMOS camera

Area of Science:

• Optics and Photonics
• Robotics and Autonomous Systems
• Computer Vision

Background:

• Underwater 3D imaging is crucial for various applications, including marine research, autonomous navigation, and underwater construction.
• Traditional methods face challenges due to light scattering, absorption, and limited visibility in aquatic environments.
• Real-time depth estimation is essential for dynamic underwater operations.

Purpose of the Study:

• To develop and validate a novel range-gated camera system for real-time underwater 3D depth estimation.
• To address the challenges posed by light scattering (forward and backward) in underwater environments.
• To achieve high-resolution and accurate depth measurements at operational speeds.

Main Methods:

• Utilized a customized fast-shutter CMOS sensor synchronized with a solid-state actively Q-switched 532 nm laser.
• Implemented a specialized depth estimation algorithm incorporating signal filtering to mitigate noise and scattering effects.
• Developed a robust peak-finding algorithm and super-resolution technique (parabolic fitting) for precise depth extraction.

Main Results:

• The system achieves real-time (10 Hz) 3D depth estimation within a typical range of 1-9 meters.
• Demonstrated depth resolution (σ) ranging from 0.8 to 9 cm, dependent on signal levels and water quality.
• Successfully estimated depth over 4.5 attenuation lengths for high albedo targets in low attenuation conditions.

Conclusions:

• The developed range-gated camera system provides a robust and accurate solution for real-time underwater 3D perception.
• The advanced algorithms effectively compensate for light scattering, enabling reliable performance in challenging aquatic conditions.
• This technology has significant potential for enhancing underwater exploration, robotics, and surveillance.