A Panoramic unwrapping approach based on log-polar mapping

Panoramic vision is widely applied in the field of computer vision, such as video - conference, Robotic navigation and Security Monitoring. Omnidirectional vision system can provide image with 360 degrees as a field of view. In many applications, its useful to un wrap the original images with distortion to panoramic images which are more adapted to human visual system. The traditional unwrapping approach is divided the original image to several areas and expansion one by one. However, this method will greatly increase the computational complexity. To overcome this problem, we present an improvement of this method, an omnidirectional image unwrapping approach based on log-polar. Experimental results indicate the unwrapping approach based on log-polar mapping is not only effective to enhance the quality of unwrapped panoramic images, but also with fast processing.