Automated Chinese medicinal plants classification based on machine learning using leaf morpho-colorimetry, fractal dimension and visible/near infrared spectroscopy

• DOI: 10.25165/j.ijabe.20191202.4637
• Publication Date: 2019-04-17T03:34:52Z
• Authors: Jinru Xue; ; Sigfredo Fuentes; Carlos Poblete-Ec...; Claudia Gonzalez ...; Eden Tongson; Hejuan Du; Baofeng Su
• Keywords: Morpho; Identification

The identification of Chinese medicinal plants was conducted to rely on ampelographic manual assessment by experts. More recently, machine learning algorithms for pattern recognition have been successfully applied leaf in other plant species. These new tools make the classification easier, more efficient and cost effective. This study showed comparative results between models obtained from two methods: i) a morpho-colorimetric method ii) visible (VIS)/Near Infrared (NIR) spectral analysis sampled leaves 20 different plants. Specifically, automated image VIS/NIR based parameters were used separately as inputs construct customized artificial neural network (ANN) models. Results that ANN model developed using (Model A) had an accuracy 98.3% studied. In case data B), averaged spectra per 92.5% all used. Model A has advantage being effective, requiring only normal document scanner measuring instrument. can be adapted non-destructive in-situ portable wireless scanners. B combines fast, advantages spectroscopy, which rapid non-invasive applications analyzing specific light overtones assess concentration pigments such chlorophyll, anthocyanins others are related active compounds Keywords: ampelography, computer vision, networks, recognition, DOI: 10.25165/j.ijabe.20191202.4637 Citation: Xue J R, Fuentes S, Poblete-Echeverria C, Viejo C G, Tongson E, Du H J, et al. Automated morpho-colorimetry, fractal dimension visible/near infrared spectroscopy. Int Agric & Biol Eng, 2019; 12(2): 123–131.