Since hyperspectral imaging has been developed as a powerful technique in earth remote sensing, this technique continues being utilized in medical, biological, agricultural, and industrial areas. This technique is capable of providing radiometric measurements over contiguous spectral ranges for every pixel of an image. Thus, data from hyperspectral images contain two–dimensional spatial information and spectral information as well. Our research group has been working on development of hyperspectral imaging technology for food quality and safety evaluation. For the detection of contaminants on broiler carcasses in poultry industry, real-time hyperspectral imaging system with line-scan platform has been developed and tested at processing line speed. In particular, hyperspectral imaging methods at microscopic level have been developed for food safety applications, because early and rapid detection of foodborne pathogen during food processing are crucial to protect foodborne illness and prevent outbreaks to the public. In this presentation, the state-of-the-art hyperspectral microscope imaging (HMI) methods to identify and classify foodborne bacteria are reported. More detail, research findings regarding development of HMI methods including immobilization of live bacterial samples on microscopic glass slide for HMI scan, dark-field illumination for scattering intensity from bacterial images at cellular level, acousto-optic tunable filters (AOTF)-based HMI image acquisition from live bacteria, HMI image processing and analysis at both spatial and spectral domains, development of classification models with deep-learning as well as various chemometric methods, and finally, early and rapid detection of foodborne pathogens with HMI technique for food industry applications are presented.
|