SUMMARY

The objective of this research is to coordinate motion control and image processing in a fast moving robotic system inspired by the observation of the human visual system. The human eye is an organ that allows the brain to recognize the environment. Because the eye has a limited field of view, the human visual system scans the environment or changes the region of interest by changing the direction of the person’s gaze. Saccade is one of the fastest and most accurate movements in the human eye. However, the human visual system receives blurry environmental information due to a finite integration time when a saccade takes place. Evidence indicates that this blurred information, called motion smear, is partially compensated by neural processes and that the human brain predicts or uses information on eye movements to reduce blurring.
Inspired by the human ocular system, this study will develop a de-blurring method using a robotic camera positioner system. A dynamics-based motion de-blurring method is introduced to restore images during rapid motion of the robotic vision system. Given mechanical properties and command inputs to the actuation system, a blur kernel can be estimated in a sensorless and fast manner while existing deblurring methods are computationally expensive and require user inputs. The developed method will be combined with parallel computing for control and image processing that enables real-time image de-blurring and quick scanning of the environment to generate a panoramic image.