Simple correction model for blurred images of uncooled bolometer type infrared cameras

摘要

Image blurring often occur when one uses uncooled bolometer type infrared cameras. We will report a simple way tocorrect the blurred images, which differs from the well-known correction method such as reconstruction of point-spreadfunction and deconvolution. Uncooled infrared sensors have a long thermal time constant which causes to overflow thesignal to the adjacent pixels. It is the reason why the images are blurred by the motion and the vibration. Therefore, theimage correction is necessary to restore the signal and the shape of the objects under that condition. Our correctionmodel enables a real-time image processing due to simple calculation method. We demonstrated the correction model byusing an infrared camera attached to a vehicle, with a gyroscopic sensor to detect the orientation and the motion of thecamera. When the camera is shaken by 15 degrees per second, our correction model recovers signal intensity by 65 % ofthe still image compared with 32 % after blurred, keeping noise level low.