ABSTRACT
This chapter describes the implementation of a methodology for the correct acquisition of hyperspectral images, with a view to subsequently ensuring the development of statistically robust models for the qualitative and quantitative characterization of meat and bone meal using Hyperspectral chemical imaging (HCI). HCI is gaining popularity as a non-destructive tool for the quality monitoring of agricultural and food products. A generic problem in HCI is the very low spatial and spectral repeatability of image cubes. Increasing attention is being paid to this issue by researchers investigating the use of HCI in remote sensing. Researchers and instrument producers have different views about how often dark and white references should be analyzed for time and image-quality optimization. During the course of images analysis, it is possible to generate infinities by dividing a scalar by 0 and undefined numbers by dividing 0 by 0.
