Authors: Rahul Raman; Sambit Bakshi; Pankaj K. Sa
Addresses: Department of Computer Science and Engineering, National Institute of Technology Rourkela, Rourkela – 769008, Odisha, India ' Department of Computer Science and Engineering, National Institute of Technology Rourkela, Rourkela – 769008, Odisha, India ' Department of Computer Science and Engineering, National Institute of Technology Rourkela, Rourkela – 769008, Odisha, India
Abstract: Past few years have seen exhaustive research in the field of camera localisation. As the era changed from single to multiple cameras, so is the paradigm shift from centralised to distributed algorithms. Euclidean geometry has been explored and the concept of Lie algebra has been touched for more subtle and non-Euclidean details. View overlaps in vision-based algorithms have been optimised and several depth measurement techniques have been implemented to extend the localisation from 2D to 3D space. LED-based techniques like triangulation and LED triangle have given depth measurement alternatives for 3D localisation whereas epipolar geometry has localised cameras with only image information. Multilateration-based approach has used anchor nodes for camera localisation whereas a few distributed algorithms (viz. DALT, DILOC) have used iterations for refinement of estimated locations. As the area under cover increased, wireless network has taken over and many algorithms have been developed for wireless networked cameras. Simultaneous existences of diverse algorithms belonging to different paradigms are needed to meet the requirement of deployment in diverse scenarios. This paper discusses the evolution from the localisation of non-camera equipped sensor network to the smart camera localisation in 3D environment that spans more than a decade.
Keywords: multi-camera localisation; camera sensor networks; surveillance; vision graph; belief propagation; multiple cameras; camera localisation; wireless networks; wireless sensor networks; WSNs.
International Journal of Machine Intelligence and Sensory Signal Processing, 2013 Vol.1 No.1, pp.91 - 109
Received: 23 Jun 2012
Accepted: 19 Aug 2012
Published online: 19 Mar 2013 *