Hyperspectral imaging (HSI) is a relatively fast analytical method that is being used for quality testing of natural foods (e.g. fruits and vegetables). This method has been proposed to replace existing manual off-line sensory quality tests of processed foods (e.g. milk powder). This article reviews the current status of HSI for monitoring natural and processed food product quality and identifies key knowledge gaps in recent literature that may be responsible for the limited success of HSI, especially for processed foods (e.g. milk powder) quality testing at the lab and industrial scales. Furthermore, various strategies to cope with challenges associated with limited HSI success are discussed. From the current literature, HSI has been used primarily as an off-line machine vision technique for quality testing to replace existing manual sensory tests. However, HSI lacks scenarios of testing and controlling product quality in real time, especially for processed foods. Little has been reported to date on generalising suitable design choices for the ‘successful’ implementation of HSI. To address this deficiency, the potential of HSI to achieve real-time quality control was examined using the milk powder production process as a case study.

高光谱成像（HSI）是一种相对快速的分析方法，已用于天然食品（例如水果和蔬菜）的质量测试。已经提出了这种方法来代替加工食品（例如奶粉）的现有手动离线感官质量测试。本文回顾了用于监控天然食品和加工食品质量的HSI的现状，并指出了近期文献中的关键知识空白，这可能是造成HSI有限成功的原因，尤其是在实验室中对加工食品（例如奶粉）进行质量测试时，工业规模。此外，讨论了应对与有限的HSI成功相关的挑战的各种策略。从目前的文献来看，HSI主要用作离线机器视觉技术进行质量测试，以代替现有的手动感官测试。但是，恒指缺乏实时测试和控制产品质量的方案，特别是对于加工食品。迄今为止，几乎没有关于为“成功”实施HSI归纳合适的设计选择的报道。为了解决这一缺陷，以奶粉生产过程为例，研究了恒指实现实时质量控制的潜力。