An efficient omnidirectional image unwrapping approach

Omnidirectional cameras are widely used almost everywhere especially in robotics and security. These kinds of cameras provide circular images that can be difficult to be interpreted by humans. Hence it must be transformed to be understandable by humans or to be treated with usual systems. This transformation is called unwrapping. The unwrapping may be time consuming and can decrease the performance of the real-time systems. Besides, the unwrapping can affect the quality of obtained images. To overcome these problems, we present an efficient parallel omnidirectional image unwrapping approach based on image partitioning. Experimental results indicate that our unwrapping approach has fast processing and gives better quality of unwrapped panoramic images.