Researchers have introduced a new vision system that takes inspiration from the eyes of cats, aimed at improving the ability to detect objects in varying lighting. The system, which has a distinct shape and reflective characteristics, minimizes glare in bright conditions while enhancing sensitivity in low-light situations. By filtering out extraneous details, this technology markedly enhances the functionality of single-lens cameras, representing a significant leap forward in robotic vision technology.
As autonomous technologies such as drones, self-driving cars, and robots become increasingly prevalent in our everyday lives, they frequently encounter challenges in “seeing” clearly in diverse environments—whether it be bright sunshine, dim lighting, or situations where objects blend into complex surroundings. Fascinatingly, nature may hold answers to these challenges.
Cats possess remarkable eyesight adaptable to both bright and dark settings. Their eyes are specially designed: during daylight, their narrow slit-like pupils help them concentrate and mitigate glare. In contrast, their pupils expand at night to capture more light, and a reflective layer known as the tapetum lucidum enhances their low-light vision, giving their eyes that signature shine.
A research team from Korea, led by Professor Young Min Song at the Gwangju Institute of Science and Technology (GIST), has developed an innovative vision system that leverages advanced lenses and sensors modeled after feline eyes. This new system is equipped with a slit-shaped aperture, akin to a cat’s vertical pupil, which aids in filtering out unnecessary light and honing in on important objects. It further incorporates a special reflective layer similar to those found in cat eyes, enhancing visibility in dim conditions. This research was published in the journal Science Advances on September 18, 2024, marking a significant progress in artificial vision technology by showcasing improved object detection and recognition, positioning it as a leader in the evolution of autonomous robotics.
“Robotic cameras often have difficulty identifying objects against busy or camouflaged backgrounds, particularly under changing light conditions. Our design addresses this issue by allowing robots to obscure irrelevant details while concentrating on key objects,” Prof. Song clarifies. Additionally, this design is energy-efficient, relying on the lens’s architecture rather than demanding extensive computational power.
This pioneering technology paves the way for exciting real-world applications, reshaping the field of robotic vision. The sophisticated vision system is set to enhance the accuracy of drones, security robots, and self-driving vehicles, enabling them to navigate complex environments and perform tasks with exceptional precision. “From search-and-rescue missions to industrial monitoring, these advanced robotic eyes are prepared to support or even substitute human endeavors in various critical situations,” emphasizes Prof. Song.
