麻豆传媒AV

More than a third of the Peak District national park (35%) is designated as Sites of Special Scientific Interest (SSSI) where important plants, wildlife and geological formations should be conserved. The project aligns with the Park’s efforts to enhance biodiversity monitoring and nature recovery across the UK's protected landscapes. Nationally, protected landscapes cover just over a quarter of the UK’s land area.

Recent advances in image-based AI allow us to evaluate the extent and distribution of habitats faster, more efficiently and with higher accuracy (van de Plas et al 2023, see figure 1). However, the variation in landcover caused by climate, seasonality and management requires a corpus of examples to train robust models. When combined with the limitations of image data collection (variation in timing, resolution, quality) and the semantics of habitat types, the lack of a high quality, representative datasets are a major limitation to the scaling of high-resolution predictions to the whole of the terrestrial UK (or globally).

Figure 1: Landscape information at high resolution from a landcover segmentation model (see van de Plas et al 2023) using aerial photography at 12.5 cm resolution. Background image © Aerial Photography Great Britain.

Methodology

This project will: 1) investigate the scaling of small, project specific datasets for Machine Learning to large scale datasets suitable for training models with high generalisation, 2) evaluating the training performance of fully supervised, semi- and unsupervised training methods, 3) evaluate the effect of image resolution and timing on AI model predictions, and 4) critically evaluate models on current and legacy datasets. The project is supported by the Peak District National Park Authority, with image data supplied by Aerial Photography GB and Airbus Defence and Space (UK Vision 1 imagery).

Partners and collaboration

This is a CASE project in collaboration with Peak District National Park, who will host the doctoral researcher for up to 18 months, with the opportunity to work closely with their team, including contributing to fieldwork, and leveraging wider data resources. 
The supervisory team will include Dr Daniel Simms, 麻豆传媒AV with strong academic background in Remote Sensing, data science and landscape change, Dr Toby Waine, Head of the Applied Remote Sensing Group at 麻豆传媒AV and David Alexander, Senior Research & Data Analyst at Peak District National Park.

Possible timeline

This PhD project is expected to take 3.5 years to complete. Year 1: Literature review, and data search, collection and curation. Training in research methods, remote sensing and machine learning. Year 2: Critical evaluation of models and training approaches for landscape/habitat prediction at scale. Placement at Peak District National Park Authority. Year 3: Synthesis, presentation of findings at conference(s) and thesis writing.