Multi digital correlated double sampling based high speed shutter-less algorithm for microbolometer

Abstract

Uncooled thermal infrared (IR) camera typically exhibits large DC offset variations and responsivity variations from pixel to pixel. Not only for these variations but PVT (Process, Voltage and Temperature) drift for time cause the undesired output. In order to get a scene information properly, non-uniformity and fixed pattern noise (FPN) from PVT drift must be corrected with reference temperature source such as blackbody. However, the calibration process is not only time-consuming and costly but also must be implemented periodically. To solve the PVT drift, especially, thermal infrared camera typically needs auxiliary component such as TEC (Thermo-Electric Cooler) and shutter during online operation but these additional components cause SWNPC (Size, Weight, Noise, Power, Cost) problems and discontinuity of the image during shutter actuation. This paper proposes novel approaches to overcome the time-consuming and costly calibration without additional components such as TEC and shutter with three ideas

MDCDS (Multi Digital Correlated Double Sampling), Pixel-synchronized temperature sensor and calibration algorithm using chamber wall. This paper includes four-type of the shutter-less algorithms as the number of blackbodies and calibration methods, compared with conventional shutter-based calibration. From this comparison, this research has come to the conclusion that proposed shutter-less algorithms have a better performance for memory, time, cost and RNU, even similar with blackbody calibration, ideal calibration method, which exhibits only random noise of the pixels.