Simulating motion blur and exposure time and evaluating its effect on image quality

Capturing images under low light conditions generally results in loss of contrast and difficulty discerning objects for both human observers and machine vision systems. To address this, the gain and exposure time are often increased to brighten the image. This may lead to the images becoming heavily affected by noise or motion blur. The impact of motion blur on image quality is therefore an important consideration. We present a simulation in which the exposure time and motion blur can be simulated and the impact on image quality metrics can be measured. Traditional image quality metrics are investigated, as well as some recently proposed alternatives. Our simulation incorporates the exposure time, motion blurring, camera setting, ambient lighting, a noise model, and optical blurring. The model allows the blurring of image quality targets and real-world images; in this paper, image quality targets are used. The variation in image quality as a function of motion and exposure time may be useful in system design, in particular, determining the sensitivity to relative motion between object and imaging system.