Hyperspectral imaging for industrial grading
I work in the Applied Technologies group working on optical sensing, imaging and machine learning. In this, I use hyperspectral imaging, multi-spectral, 3D image data, and conventional imaging to solve industrial sensing problems. The research relates to accurate classification, counting, grading, and measurement across agritech, horticulture, fisheries, forestry, fashion, and sports.
My experience with novel optical techniques, biomedical optics, customised illumination, and machine vision combines with state-of-the art machine-learning (Deep Learning / AI), and state-of-the-art edge processing and IOT technology to deliver bespoke solutions for industrial sensing.
Hyperspectral cameras can capture images of an object or scene at hundreds of different wavelengths, allowing for the detection of subtle differences in colour and reflectance that can indicate variations in material composition, chemical content, and physical properties. Hyperspectral cameras have many applications. We have used ours for specifying vision systems, to detect contaminants and defects in food, and for measuring the chemical concentrations of cannabis.