Emerging issues related to food quality include the need to ensure food safety, detect adulteration and enable traceability throughout the food supply chain. Such issues must be addressed to ensure the quality, hygiene and nutritional value of food, and thereby safeguard the health and interests of consumers. Hyperspectral imaging, which combines spectroscopic data on the chemical constituents of a sample and high-resolution imaging of its physical features, is already being used in the food industry for quality inspection purposes. However, the complexity, high cost and large size of hyperspectral imaging equipment is a substantial obstacle to the widespread implementation of this technology. Moreover, the very large, information-rich datasets generated by hyperspectral imaging are difficult to interpret appropriately. This Review describes currently available types of hyperspectral imaging hardware as well as the wide range of image analysis and data modelling tools used to analyse hyperspectral data. Illustrative examples of hyperspectral imaging applications used for food quality inspection are described in detail, and future developments in hyperspectral imaging are presented. The overall aim of this Review is to provide guidance for non-specialist researchers in the selection of hyperspectral imaging equipment, software and models that are appropriate for their intended application.
Rational and scientific use of hyperspectral imaging involves the selection of appropriate imaging hardware and data analysis software. Sun et al. describe applications of hyperspectral imaging in food quality inspection and provide guidance for non-specialist researchers aiming to implement this technology.
Key points: Appropriate selection of hardware components of hyperspectral imaging systems is important because they determine the quality of the data obtained.Image analysis and modelling tools can help researchers mine useful information from information-rich hyperspectral datasets.Future development of hyperspectral imaging applications is likely to involve device miniaturization, improved analysis methods, data standardization and data-sharing initiatives.Further expansion of hyperspectral imaging applications is expected.