Real-Time Hyperspectral Imaging for Food Safety

Abstract

Multispectral imaging systems with selected bands can commonly be used for real-time applications of food processing. Recent research has demonstrated several image processing methods including binning, noise removal filter, and appropriate morphological analysis in real-time mode can remove most false positive/negative errors. Recently researchers developed a real-time hyperspectral imaging platform that has the ability for multi-tasking during food processes. Such a system can be installed at different locations on a food processing line to solve significant food safety problems such as disease and contaminant detection simultaneously. Real-time food inspection with a developed line-scan hyperspectral imaging system is possible, with abilities of data binning utilizing a random track mode of the EMCCD sensor, and the custom software supporting multitasks. The recent development of a line-scan hyperspectral imaging system for real-time multispectral imaging applications for the food industry is demonstrated. The real-time hyperspectral imaging system consists of a spectrograph, EMCCD camera, and real-time image processing software. The imaging system can be easily modified for other real-time food inspection applications, with simple parameter changes in the software during processing. The real-time image processing software architecture is based on the ping pong memory, and a circular buffer for the multitasking of image grabbing and processing. An image-based internal triggering (i.e. polling) algorithm is developed to determine the start and end positions of objects with Microsoft Visual C++ environment.