A novel hyperspectral microscope imaging technology for rapid evaluation of particle size distribution in matcha. (May 2020)

• Record Type: Journal Article
• Authors: Ouyang, Qin; Yang, Yongcun; Park, Bosoon; Kang, Rui; Wu, Jizhong; Chen, Quansheng; Guo, Zhiming; Li, Huanhuan
• Is Part Of: Journal of food engineering. Volume 272(2020)
• Publication Date: 2020-05
• Subjects: Hyperspectral microscope imaging -- Matcha -- Particle size -- Data fusion -- Multivariate analysis -- Fine tea powder; Food industry and trade -- Periodicals; Food -- Analysis -- Periodicals; Aliments -- Industrie et commerce -- Périodiques; Aliments -- Analyse -- Périodiques; Aliments -- Recherche -- Périodiques; 664.005
• DOI: 10.1016/j.jfoodeng.2019.109782
• Languages: English
• ISSNs: 0260-8774

Abstract

Abstract : Hyperspectral microscope imaging (HMI) technology as a novel approach was proposed to evaluate physical characteristics of matcha. Particle size distribution as one of the significant physical characteristics was investigated. Data fusion which integrated of textural features from images at 524 nm and key spectral features selected by competitive adaptive reweighed sampling (CARS) were as the raw data for modeling. Models were optimized by cross-validation. Results showed that the performance of models was improved with data fusion. The best ANN models with data fusion were achieved with R p (correlation coefficient in prediction set) of 0.8020 for D10, 0.8414 for D20, 0.8238 for D30, 0.8124 for D40, 0.8058 for D50, 0.8157 for D60, 0.7643 for D70, 0.7360 for D80, and 0.6313 for D90, respectively. This work demonstrated that HMI technology as a rapid, accurate and high effective protocol has great potential in predicting particle size distribution in matcha powder. Highlights: Particle size distribution in matcha was creatively assessed using HMI technology. CARS selected characteristic spectral variables to simplify and enhance models. Data fusion models combining image with spectra information can improve performance.