Why trust Greenbot
Robots may be getting smarter, but their vision is often far from perfect. Meanwhile, cats excel in these situations, thanks to eyes designed for focus and night vision.
That being said, inspired by these feline features, researchers at the Gwangju Institute of Science and Technology (GIST) have crafted a new robotic vision system that mimics a cat’s vision. The South Korean researchers’ findings were published in Science Advances on September 18.
This system allows robots to see and understand their surroundings in real-time. It helps them make faster and more accurate decisions, even in tricky lighting.
“Robotic cameras often struggle to spot objects in busy or camouflaged backgrounds, especially when lighting conditions change,” said lead researcher Professor Young Min Song. “Our design solves this by letting robots blur out unnecessary details and focus on important objects.”
Cats have better vision than humans due to their vertically elongated pupils and a tapetum lucidum. Their adaptive vision enhances focus and reflects light to improve low-light visibility.
The researchers’ design includes a slit-like elliptical aperture to filter excess light and a silicon photodiode array backed by patterned metal reflectors to mimic the tapetum lucidum. This combination allows for better light absorption and enhanced vision in low-light scenarios.
Advantages Over Traditional Systems
Traditional systems often rely on computer algorithms to process images. Whereas, this cat-inspired approach utilizes optimized hardware to innovate the system. This reduces computational demands, making it more energy-efficient. This makes it a significant advantage for drones and other autonomous robots.
During tests, the system proved capable of maintaining a sharp focus on targets, even when the background was cluttered or camouflaged. This was due to the system’s unique ability to blur the background while concentrating on the main object. In contrast, conventional systems struggled, especially under bright light.
Real-World Applications and Future Challenges
With this vision system, machines could better identify, track, and recognize objects in dynamic conditions where human eyes might fail.
“From search-and-rescue operations to industrial monitoring, these cutting-edge robotic eyes stand ready to complement or even replace human efforts in a variety of critical scenarios,” Professor Song emphasized.
While the system is effective, there are limitations to overcome. One notable challenge is its narrow field of view. The team suggests integrating movements similar to a cat’s head to expand this field, making it more practical for real-world applications.
