This thesis describes a generic approach for automated land cover monitoring. Knowledge about land cover is acquired through a knowledge acquisition exercise and used to augment image analysis in order to determine land cover change direction. It is demonstrated that taking a task-oriented approach to the change problem avoids the specificity of more traditional data-oriented approaches. The approach described here involves four key analyses that have contributed to the overall problem solution: Identifying the knowledge used to determine different land cover elements (elicitation and modelling); Land cover remote sensing characteristics; Land cover bio-geographic characteristics; Investigation of the most suitable approach for combining evidence. The results of these investigations were applied to semi-natural change problems: evidence about areas known to have changed was reasoned with and change directions identified by applying the knowledge and interpreter rules of thumb. The results were compared with field surveys of the change areas, and were shown to have correctly identified the land cover change direction in each case.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:248628 |
| Date | January 2002 |
| Creators | Comber, Alexis |
| Publisher | University of Aberdeen |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0022 seconds