Document Type


Degree Name

Doctor of Philosophy (PhD)


Geography & Environmental Studies

Program Name/Specialization



Faculty of Arts

First Advisor

Dr. Colin Robertson

Advisor Role

Provide funding to support research, recommend research methods and review dissertation


Human-induced disturbances of terrestrial and aquatic ecosystems continue at alarming rates. With the advent of both raw sensor and analysis-ready datasets, the need to monitor ecosystem disturbances is now more imperative than ever; yet the task is becoming increasingly complex with increasing sources and varieties of earth observation data. In this research, computer vision methods and tools are interrogated to understand their capability for comparing spatial patterns. A critical survey of literature provides evidence that computer vision methods are relatively robust to scale and highlights issues involved in parameterization of computer vision models for characterizing significant pattern information in a geographic context. Utilizing two widely used pattern indices to compare spatial patterns in simulated and real-world datasets revealed their potential to detect subtle changes in spatial patterns which would not otherwise be feasible using traditional pixel-level techniques. A texture-based CNN model was developed to extract spatially relevant information for landscape similarity comparison; the CNN feature maps proved to be effective in distinguishing agriculture landscapes from other landscape types (e.g., forest and mountainous landscapes). For real-world human disturbance monitoring, a U-Net CNN was developed and compared with a random forest model. Both modeling frameworks exhibit promising potential to map placer mining disturbance; however, random forests proved simple to train and deploy for placer mapping, while the U-Net may be used to augment RF as it is capable of reducing misclassification errors and will benefit from increasing availability of detailed training data.

Convocation Year


Convocation Season