Growth identification of Aspergillus flavus and Aspergillus parasiticus by visible/near infrared hyperspectral imaging

Authors: CHU, XUAN - China Agriculture University; WANG, WEI - China Agriculture University; Ni, Xinzhi; ZHENG, HAITAO - China Agricultural University; ZHAO, XIN - China Agriculture University; ZHANG, REN - Tarim University; LI, YUFENG - Chinese Academy Of Sciences

Submitted to: Applied Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/12/2018
Publication Date: 3/28/2018
Citation: Chu, X., Wang, W., Ni, X., Zheng, H., Zhao, X., Zhang, R., Li, Y. 2018. Growth identification of Aspergillus flavus and Aspergillus parasiticus by visible/near infrared hyperspectral imaging. Applied Sciences. 8:Article 513.

Interpretive Summary: Use of Visible/near-infrared hyperspectral imaging (between 400 and 1000 nm wavelengths) to identify the growing process of Aspergillus flavus and Aspergillus parasiticus was evaluated. The hyperspectral images of the two fungi cultured on rose bengal medium were recorded daily for 6 d. A band ratio using two bands at 446 nm and 460 nm separated A. flavus and A. parasiticus on day 1 from the other days. Image at band of 520 nm classified A. parasiticus on day 6. However, the images for A. flavus on days 3 and 4 and A. parasiticus on days 2 and 3 overlapped. After the average spectra of each fungus in each growth day were extracted, the models built using principal components could identify fungal growth days with accuracies of 92.59% and 100% for A. flavus and A. parasiticus individually. In order to simplify the prediction models, a new sampling method was used to choose optimal wavelengths. As the results, nine (that is, 402, 442, 487, 502, 524, 553, 646, 671, and 760 nm) and seven (that is, 461, 538, 542, 742, 753, 756, and 919 nm) wavelengths were selected for A. flavus and A. parasiticus, respectively. New optimal wavelengths of the models were built, and the identification accuracies were 83.33% and 98.15% for A. flavus and A. parasiticus, respectively.

Technical Abstract: Visible/near-infrared (Vis/NIR) hyperspectral imaging (400–1000 nm) was applied to identify the growing process of Aspergillus flavus and Aspergillus parasiticus. The hyperspectral images of the two fungi cultured on rose bengal medium were recorded daily for 6 d. A band ratio using two bands at 446 nm and 460 nm separated A. flavus and A. parasiticus on day 1 from other days. Image at band of 520 nm classified A. parasiticus on day 6. Principal component analysis (PCA) was performed on the cleaned hyperspectral images. The score plot of the second to sixth principal components (i.e., PC2 to PC6) gave a rough clustering of fungi in the same incubation time. However, in the plot, A. flavus on day 3 and day 4 and A. parasiticus on day 2 and day 3 overlapped. The average spectra of each fungus in each growth day were extracted, and then PCA and support vector machine (SVM) classifier were applied to the full spectral range. SVM models built by PC2 to PC6 could identify fungal growth days with accuracies of 92.59% and 100% for A. flavus and A. parasiticus individually. In order to simplify the prediction models, competitive adaptive reweighted sampling (CARS) was employed to choose optimal wavelengths. As the results, nine (i.e., 402, 442, 487, 502, 524, 553, 646, 671, and 760 nm) and seven (i.e., 461, 538, 542, 742, 753, 756, and 919 nm) wavelengths were selected for A. flavus and A. parasiticus, respectively. New optimal wavelengths of the SVM models were built, and the identification accuracies were 83.33% and 98.15% for A. flavus and A. parasiticus, respectively. Finally, the visualized prediction images for A. flavus and A. parasiticus in different growth days were made by applying the optimal wavelength’s SVM models on every pixel of the hyperspectral image.