Early warning system for infective sea lice: Development of holographic imaging for in-situ monitoring (RD201)

Project Outcome:

The project developed and tested an innovative method for identifying sea lice larvae using holographic imaging and artificial intelligence (AI), offering a more accurate and efficient approach to detecting and quantifying sea lice larvae in marine environments. 

Key achievements include the development of a holographic imaging system for capturing images of sea lice larvae both in laboratory settings and during field deployments. A sea lice hatchery was established to ensure a consistent supply of larvae for experimentation, enabling the creation of extensive holographic image catalogues of sea lice larvae at various developmental stages alongside other plankton species. To complement this, imaging chambers and flumes were designed and built to capture larvae images in different orientations and under different conditions (live and fixed sea lice larvae), and to quantify their abundance in the water, facilitating the development of an AI-based image classification tool using a Convolutional Neural Network (CNN) to distinguish sea lice larvae from other microorganisms to a very high accuracy. 

The methodology was successfully tested in Loch Linnhe. Collected samples of zooplankton from the field resulted in a valuable dataset of presence and distribution of sea lice larvae, revealing unexpectedly high larvae abundances in deeper, more saline waters. This dataset was achieved through traditional methods of microscopic identification by taxonomists and provided a rigorous test of the holographic imaging and AI quantification system. 

By integrating advanced technology, rigorous laboratory and field testing, and collaborative research, this work represents a significant step forward in sustainable aquaculture practices.