PhD Studentship: Multispectral Imaging and AI for Marine Litter Detection

This PhD project aims to improve the automated detection, classification, and quantification of marine litter using multispectral and hyperspectral imaging. Based at the University of East Anglia's Colour & Imaging Laboratory, the successful candidate will collaborate with the Centre for Environment, Fisheries and Aquaculture Science (CEFAS) to enhance existing deep learning models. The research involves characterising material reflectance, curating imaging datasets, and developing robust machine-learning approaches that account for varying environmental conditions. The project offers comprehensive training in spectral imaging, computer vision, and geospatial analysis, with potential opportunities for drone-based fieldwork.

• Characterise representative marine litter materials using laboratory imaging systems.
• Curate and manage large-scale imaging datasets for model training.
• Develop and refine machine-learning models for material detection and classification.
• Investigate model performance under varying illumination and sensor conditions.
• Apply techniques such as transfer learning and domain adaptation to improve model robustness.
• Evaluate the translation of laboratory models into practical drone-based monitoring systems.
• Collaborate with scientists and engineers at CEFAS to ensure research relevance.