This research focuses on the development of a novel image super-resolution algorithm for coordinate measurement in manufacturing. The main features of the algorithm are that it is fast, flexible and fully automatic. A fast algorithm is required because image-super resolution is a procedure that handles a large amount of data. Having a slow or highly complex algorithm may result in computational infeasibility. A flexible algorithm means the algorithm can be customised to handle specific problems, i.e. the algorithm can be augmented with multiple constraints and still obtain an optimal solution. This is desirable as most image super-resolution problems are specific and having the capacity to augment multiple constraints reduces the search space, thus leading to faster convergence. An automatic algorithm is viewed as ideal as it has minimum human intervention and will generate super-resolution images automatically when measured frames are input. / This study considers three issues related the developing the algorithms: the model of image super-resolution; the formulation of a flexible algorithm that is capable of augmenting multiple constraints into the model and produces optimal super-resolution images; and the optimisation technique to solve the problem formulated to ensure that the computational complexity is low. / Thesis (PhDEngineering)--University of South Australia, 2005.
| Identifer | oai:union.ndltd.org:ADTP/267282 |
| Creators | Ling, Dennis Sie Hieng. |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | copyright under review |
Page generated in 0.0019 seconds