Research interests
keywords: remote sensing, classification, segmentation, change detection, texture, uncertainty, fuzzy, visualisation
Satellite images provide a rich source of information for studying the Earth surface. One of my particular interests is in the application of pattern recognition algorithms to remote sensing. Pattern recognition techniques aim to simulate human vision to extract features from satellite images; examples include vegetation and landform classification. In mapping natural features, we often tend to draw sharp or discrete boundaries between classes. This is often unrealistic as transition zones between classes (or ecotones) often occur in natural areas. My research focuses on using fuzzy techniques to quantify uncertainty related to these transition zones. My current area of interest is in using these techniques to map vegetation and vegetation change on sub-Antarctic Heard Island and Macquarie Island.
Additionally, I am interested in using advanced and interactive visualisation techniques to provide more insight into patterns in a satellite image. Visualisation tools can help to explain the results of a complex pattern recognition algorithm in a more accessible and insightful way. The field of remote sensing is exciting to work in as technology is changing rapidly requiring development of new techniques and tools for analysis and information extraction.
More recently, I have started to develop a remote sensing system based on an unmanned aerial vehicle (UAV) for super high resolution multispectral imaging and airborne laser scanning (LiDAR). More detail about my projects can be found here.
