A distortion correction algorithm for fish-eye panoramic image of master-slave visual surveillance system

A master-slave visual surveillance system is composed of one fish-eye panoramic camera and one dynamic pan-tilt-zoom (PTZ) dome camera. In order to make PTZ dome camera zoom into all targets of interest in panoramic image, the fish-eye panoramic camera is fixed inclining towards the gravity direction, which may cause more obvious distortion. This paper proposed a novel method for the distortion correction of captured panoramic image based on the midpoint circle algorithm (MCA). The method uses incremental calculation of decision parameters to determine the pixel positions along a circle circumference, and both of the vertical and horizontal are rectilinearised. Experimental results show that our correction method based on MCA is efficient and effective. In particular, due to its low computational cost, our method can be applied on embedded camera platform without any extra hardware resources.