Thermal Infrared
for Robot Vision
DESCRIPTION
Thermal cameras enable robots to see in the complete absence of illumination by imaging the “heat” radiated from a scene. Unlike a human eye that needs reflected light, these sensors depend primarily on emitted radiation, which eliminates shadows and reduces lighting artifacts. The longer wavelengths used by these cameras also allow them to see better through visual obscurants. Despite their potential benefits, microbolometers (the affordable type of thermal camera) suffer from significant motion blur that impedes their wider adoption in robotics. We study the origins of motion blur in microbolometers by starting from their physics of image formation. We use this to adapt vision algorithms so that fast moving mobile robots can “see” in the dark.
Related Publications
M. Srinivasan Ramanagopal, Z. Zhang, R. Vasudevan, and M. Johnson-Roberson, “Pixel-Wise Motion Deblurring of Thermal Videos,” in Proceedings of Robotics: Science and Systems, Corvalis, Oregon, USA, July 2020. [website]