Snapshot Multispectral and Hyperspectral Data Processing for Estimating Food Quality Parameters

Abstract

Multispectral and hyperspectral imaging systems have proven capabilities in estimating critical food quality parameters for a number of food processing and inspection tasks. In this paper, we have developed a processing pipeline for multispectral and hyperspectral snapshot video sensors, towards detecting certain critical quality parameters like freshness, spoilage levels and storage temperatures. In particular, a set of pre-processing modules are detecting clear meat or salad observations and then a classification algorithm is responsible for detecting and labelling accordingly each pixel or sample. The experimental results and performed quantitative validation indicate the quite promising potentials of the developed approach.