Recent research by the National Trust has used Machine Learning to map changes in 20th century configurations of UK habitats such as orchards, linear features and trees across landscapes. This PhD project will extend these approaches to a wider set of habitats & landscape elements for these and earlier periods i.e. prior to systematic national mapping. It will also explore augmentation with a heterogeneous range of other datasets (such as lidar, land cover mapping) and/or combining the results with existing spatial data (such as biological survey records or Historic Landscape Characterisation maps (HLC)). Understanding the evolution of landscape features is important to help guide current and future biodiversity renovation and contextualise this with historic and cultural knowledge. Research challenges include use of computer vision to identify symbols on historic mapping, referencing map features against geospatial features, and combining disparate datasets to foster a clearer picture of past landscapes. The student will acquire advanced data science and AI skills in the context of environmental change and ecology. Within the broad area of the project, there is scope for the student to direct their research in different directions, depending on their interests and the results of early explorations.
Plus d’informations :
[Website University of Exeter]