Date Published: July 13, 2017
Publisher: Public Library of Science
Author(s): Wenxin Yin, Chu Zhang, Hongyan Zhu, Yanru Zhao, Yong He, George-John Nychas.
Near-infrared (874–1734 nm) hyperspectral imaging (NIR-HSI) technique combined with chemometric methods was used to trace origins of 1200 Chinese wolfberry samples, which from Ningxia, Inner Mongolia, Sinkiang and Qinghai in China. Two approaches, named pixel-wise and object-wise, were investigated to discriminative the origin of these Chinese wolfberries. The pixel-wise classification assigned a class to each pixel from individual Chinese wolfberries, and with this approach, the differences in the Chinese wolfberries from four origins were reflected intuitively. Object-wise classification was performed using mean spectra. The average spectral information of all pixels of each sample in the hyperspectral image was extracted as the representative spectrum of a sample, and then discriminant analysis models of the origins of Chinese wolfberries were established based on these average spectra. Specifically, the spectral curves of all samples were collected, and after removal of obvious noise, the spectra of 972–1609 nm were viewed as the spectra of wolfberry. Then, the spectral curves were pretreated with moving average smoothing (MA), and discriminant analysis models including support vector machine (SVM), neural network with radial basis function (NN-RBF) and extreme learning machine (ELM) were established based on the full-band spectra, the extracted characteristic wavelengths from loadings of principal component analysis (PCA) and 2nd derivative spectra, respectively. Among these models, the recognition accuracies of the calibration set and prediction set of the ELM model based on extracted characteristic wavelengths from loadings of PCA were higher than 90%. The model not only ensured a high recognition rate but also simplified the model and was conducive to future rapid on-line testing. The results revealed that NIR-HSI combined with PCA loadings-ELM could rapidly trace the origins of Chinese wolfberries.
Chinese wolfberry is a multi-branched shrub in the family Solanaceae, and the fruit, skin, and leaves can be used as medicine . What’s more, the wolfberry shrubs are widely planted in Inner Mongolia, Shaanxi, Gansu, Ningxia, Qinghai and Sinkiang and other places in China for it has excellent soil and water conservation capacity . It is well accepted that the growing environment may alter the chemical composition and biological properties of a selected botanical . Most consumers favor the Ningxia wolfberries, which have a characteristic large fruit, nice shape, high content of active ingredient and a wide range of medicinal value . However, with the frequent mixing of fruits from different origins in the market in recent years, the quality of Ningxia wolfberries is difficult to guarantee. According to most researches, the geographical origins of Chinese wolfberries can be identified based on observing the shape of the wolfberries and using chemical methods to detect internal quality, however, these methods are time-consuming, destructive to the samples and with low detection accuracy [4–5]. Therefore, establishing rapid, nondestructive and high-accurate methods to trace the origin of Chinese wolfberries is urgent. Meanwhile, these analytical methods are also needed for wolfberry breeding efforts to obtain improved cultivars with enhanced nutritional and nutraceutical quality and farm gate value for commercial production of Ningxia wolfberries .
The origins of Chinese wolfberries were traced using an NIR-HSI system combined with extracted characteristic bands and different discriminant analysis models. From the perspective of the pixel spectra of the Chinese wolfberries combined with the spatial distribution of the Chinese wolfberries, a principal component pseudo-color map was drawn, and the differences of wolfberries from four origins were displayed intuitively. From the perspective of wolfberry samples, different discriminant analysis models were built on the full spectra and the characteristic wavelengths extracted by PCA loadings and 2nd derivative spectra. Following analysis and comparison, the discriminant models based on the full spectra were better than those based on the characteristic wavelengths. Among the discriminant analysis modeling methods, ELM algorithm obtained the best discriminant effects. ELM model based on the characteristic wavelengths extracted by PCA loadings not only provided high recognition accuracy but also simplified the model, which facilitated rapid on-line detection. In future research, as many origins as possible of Chinese wolfberries should be studied to establish a more robust and wider range of identifications in models of the origins of Chinese wolfberries, and the feasibility should be studied of applying HSI technique to detect quality in Chinese wolfberry and determine whether Chinese wolfberry has been artificially smoked.