Seeing through murky waters
MIT and WHOI researchers developed a sonar-camera system that lets underwater robots navigate sediment clouds in real time, without waiting for the water to clear.
Researchers at MIT and the Woods Hole Oceanographic Institution have built a navigation system that allows remotely operated underwater vehicles to map and move through sediment-clouded water without stopping to wait for visibility to return, according to MIT Technology Review.
The system was developed by Amy Phung, SM '23, PhD '26, and her advisor Richard Camilli, SM '00, PhD '03, of WHOI. When an underwater vehicle settles on the seafloor or digs through a sand bed, it can kick up dense clouds of sediment that defeat onboard cameras. The new approach sidesteps that problem by first using sonar to rapidly map the vehicle's surroundings — a technology that works equally well in cloudy and clear water — and then guiding the vehicle close enough to specific objects for cameras to capture useful detail.
To make the mapping fast enough for real-time use, the team combined the sonar system with an image-matching algorithm developed by researchers in France. That algorithm quickly estimates the relative depth of each pixel in a two-dimensional scene, per MIT Technology Review, enabling the vehicle to build a workable picture of its environment on the fly rather than relying on post-processing.
Camilli described the challenge in an analogy: "An analogy would be if you were to go into a china shop in the dark, and try to pick your way around to find a specific coffee mug without knocking things over. This would allow you to do that." He and Phung say potential applications range from scientific exploration and underwater construction and maintenance to the handling of unexploded undersea mines.
The work represents a practical step toward more autonomous underwater operations in conditions where optical sensors alone consistently fail. By pairing low-resolution but penetrating sonar with a depth-estimation algorithm for camera-based close-up inspection, the system avoids the binary choice operators currently face: either halt until sediment clears or risk damaging equipment or objectives by navigating blind.
No comments yet — start the thread.