University of Maryland Scientists Revolutionize Robot Vision with New Camera System

To mimic microsaccades, the team used a spinning prism inside the AMI-EV to reroute light beams recorded by the lens

Scientists from the University of Maryland have developed an innovative camera system that significantly enhances robots' vision and response to their surroundings. Drawing inspiration from the human eye, this breakthrough technology mimics the small involuntary movements our eyes make to maintain stable and clear vision over time. The researchers introduced the Artificial Microsaccade-Enhanced Event Camera (AMI-EV), with their findings recently published in Science Robotics. The key to this advancement is the replication of microsaccades—brief, rapid eye movements that occur unintentionally when focusing on a specific field of vision. These tiny, constant motions enable the human eye to maintain precise focus on an object, capturing its visual textures, including color, depth, and shadowing.

To mimic microsaccades, the team used a spinning prism inside the AMI-EV to reroute light beams recorded by the lens. This continuous rotation of the prism replicated the human eye's movements, allowing the camera to stabilize an object's textures similarly. Initial tests of the AMI-EV showed remarkable precision in capturing and displaying movement across various scenarios, such as identifying fast-moving shapes and detecting human pulses. Moreover, the AMI-EV can record motion at tens of thousands of frames per second, far exceeding the typical 30 to 1000 frames per second of most commercial cameras. This capability offers a more realistic and fluid portrayal of motion, potentially transforming fields ranging from security monitoring to augmented reality experiences and enhancing astronomical imaging.