Fusion of 3D models constructed on the basis of omnidirectional images

Building 3D models of the scene is a key task in computer vision, new and very dynamic developing discipline of computer science. Spatial models of real objects and indoors can be a target itself (for example for the purpose of multimedia applications), or just an intermediate stage (for instance in navigation or control processes). In this work the contribution of omnidirectional images to 3D modeling is shown.