研究目的
To provide a comprehensive review of hyperspectral techniques and chemometric methods for assessing quality parameters of roots and tubers, including physical properties, chemical constituents, varietal authentication, gradation aspects, and defect aspects.
研究成果
Hyperspectral techniques, combined with chemometric methods, offer rapid, non-invasive, and accurate quality assessment of roots and tubers. They show promise for evaluating physical, chemical, varietal, grading, and defect parameters, with NIR being the most used spectral range and PLSR the most common modeling approach. Future work should focus on improving model accuracy, optimizing data processing, and developing high-speed hardware for online applications.
研究不足
The review highlights challenges such as limited light penetration depth in tuber samples, interference from high water content, large data size requiring significant storage and computation, need for prior knowledge from reference methods, and time-consuming modeling processes. It also notes the difficulty in real-time application due to hardware and cost constraints.
1：Experimental Design and Method Selection:
The paper is a review article, not an experimental study, so it does not describe a specific experimental design or methodology. It summarizes existing literature on hyperspectral techniques (spectroscopy and hyperspectral imaging) combined with chemometric analyses for quality evaluation of roots and tubers.
2：Sample Selection and Data Sources:
The review covers various studies involving samples such as potato, sweet potato, cassava, yam, taro, and sugar beet, but no specific sample selection criteria or data sources are detailed.
3：List of Experimental Equipment and Materials:
Not applicable as it is a review; specific equipment from cited studies are mentioned in the references but not listed here.
4：Experimental Procedures and Operational Workflow:
Not applicable; the review discusses general procedures like spectral data acquisition and multivariate analysis without step-by-step details.
5：Data Analysis Methods:
Chemometric methods such as partial least squares regression (PLSR), artificial neural networks (ANN), principal component regression (PCR), support vector machines (SVM), multiple linear regression (MLR), and least square support vector machine (LS-SVM) are mentioned, along with variable selection techniques like genetic algorithm (GA) and successive projection algorithm (SPA).
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
