No description available.

No description available.

No description available.

No description available.

This thesis describes research on the use of Java to develop cross-platform computer vision applications for mobile phones with integrated cameras. The particular area of research that we are interested in is Mobile Augmented Reality (AR). Currently there is no computer vision library which can be used for mobile Augmented Reality using the J2ME platform. This thesis introduces the structure of our J2ME computer vision library and describes the implementation of algorithms in our library. We also present several sample applications on J2ME enabled mobile phones and report on experiments conducted to evaluate the compatibility, portability and efficiency of the implemented algorithms.

Image based scene reconstruction from multiple views is an interesting challenge, with many ambiguities and sources of noise. One approach to scene reconstruction is Voxel Colouring, Seitz and Dyer [26], which uses colour information in images and handles the problem of occlusion. Culbertson and Malzbender [11], introduced Generalised Voxel Colouring (GVC) which uses projection and rasterization to establish global scene visibility. Our work has involved investigating the use of ray traversal as an efficient alternative. We have developed two main approaches along this line, Ray Images and Ray Buckets. Comparisons between implementations of our algorithms and variations of GVC are presented, as well as applications to areas of optimisation colour consistency and level of detail. Ray traversal seems a promising approach to scene visibility, but requires more work to be of practical use. Our methods show some advantages over existing approaches in time use. However we have not been as succesful as an- ticipated in reconstruction quality shown by implementation of optimisation colour consistency.

A survey of robotized seam-tracking techniques was conducted in preparation for a project consisting of using an infrared camera on a robot for on one side collect continuous weld images for NDT inspection and on the other one track the weld joint. The tracking system can be used to discover the position of the weld bead without previous information. The robust system is outlined, along with its merits and disadvantages.

No description available.

No description available.

The non-rigid structure from motion problem typically involves recovering the 3D trajectories of a set of scene points, from their corresponding image trajectories. In this thesis, the assumption of locally-rigid motion is used to regularize this otherwise underconstrained problem. The key idea is that even when a scene undergoes complex global deformations, the trajectories of local triplets of scene points can often be approximated by the vertices of a rigidly moving triangle. This intuition informs our bottom-up reconstruction procedure, which discovers such triplets through a hypothesis and test framework. To this end, a rigid triangle model is fit to the proposed image trajectories and evaluated using a procedure that we call 3-SFM. The recovered triangle models are then integrated into a global solution, by resolving their orthographic depth ip and translation ambiguities. Lastly, we consider using this solution to initialize an energy based model, subject to a set of soft isometric constraints, in order to allow each observation to constrain the global scene structure. Results on several sequences, both our own and from related work, suggest that these models are applicable in diverse and challenging scenes, such as those including multiple deforming bodies.