Detection Models Study of Chlorophyll in Winter Wheat’s Leaves by Reflectance Spectra and Artificial Neural Networks
Reflectance spectra of living wheat’s leaves were acquired by portable radiation spectrometer. Multivariate scatter correction (MSC) method was used to preprocess the spectra, and quantitative analysis detection models were built up by artificial neural networks (ANN). The models of chlorophyll, chlorophyll a, and chlorophyll b in leaves were built up and tested. In calibration set and predicted set, for chlorophyll in leaves, the correlation coefficient (R) were 0.927 and 0.818, the standard deviation (SD) were 0.252 and 0.470; for chlorophyll a in leaves, the correlation coefficient (R) were 0.780 and 0.627, the standard deviation (SD) were 0.441 and 0.592; for chlorophyll b in leaves, the correlation coefficient (R) were 0.876 and 0.871, the standard deviation (SD) were 0.205 and 0.259. The results show that, using portable radiation spectrometer to acquire the reflectance spectra, multivariate scatter correction to preprocess the spectra, and artificial neural networks to build up the models, the chlorophyll, chlorophyll a and chlorophyll b in winter wheat’s leaves may be quantitative detected.
KeywordsArtificial neural networks Spectral analysis Winter wheat’s leaves Chlorophyll Chlorophyll a Chlorophyll b
The authors acknowledge the National Key Research and Development Program (Grant No. 2016YFD0200602).
- 1.Fan, Z., Simone, G., Yuping, B., et al.: Estimation of wheat water status by reflectance measurements. Trans. CSAE 21(S), 215–217 (2005)Google Scholar
- 2.Ji, H., Wang, P., Yan, T.: Estimations of chlorophyll and water contents in live leaf of winter wheat with reflectance spectroscopy. Spectrosc. Spectr. Anal. 27(3), 514–516 (2007)Google Scholar
- 3.Li, F., Zhao, C., Wang, J., et al.: Diagnosis of nitrogen nutrition of flue-cured tobacco with chlorophyll meter. Plant Nutr. Fertil. Sci. 13(1), 136–142 (2007)Google Scholar
- 4.Wang, D., Ji, J., Gao, H.: The effect of MSC spectral pretreatment regions on near infrared spectroscopy calibration results. Spectrosc. Spectr. Anal. 34(9), 2387–2390 (2014)Google Scholar
- 5.Yu, D., Cheng, W., Wang, Q., et al.: Construction of analysis model of rice seed components based on near infrared reflectance spectroscopy. J. Light. Scatt. 27(4), 384–389 (2015)Google Scholar
- 6.Li, L., Huang, H., Zhao, S., et al.: NIR spectra detection model of protein, fat, total sugar and moisture in rice. J. Chin. Cereals Oils Assoc. 32(7), 121–126 (2017)Google Scholar
- 7.Cao, H., Li, D., Liu, L., et al.: Near infrared spectroscopy quantitative analysis model based on incremental neural network with partial least squares. Spectrosc. Spectr. Anal. 34(10), 2799–2803 (2014)Google Scholar
- 8.Lai, L., Ma, W., Chen, H.: Wheat protein nondestructive analysis with near infrared reflectance spectroscopy combined with artificial neutral networks. J. China Univ. Metrol. 26(1), 55–59 (2015)Google Scholar
- 9.Kuang, J., Guan, X., Liu, J.: Rapid determination of protein and fat contents in raw milk by near infrared spectroscopy analysis. J. Anal. Sci. 31(6), 783–786 (2015)Google Scholar
- 10.Zheng, W., Ming, J., Yang, M., et al.: Hyperspectral estimation of rice pigment content based on band depth analysis and BP neural network. Chin. J. Eco-Agric. 25(8), 1224–1235 (2017)Google Scholar
- 11.Yao, X., Wang, X., Huang, Y., et al.: Estimation of sugar to nitrogen ratio in wheat leaves with near infrared spectrometry. Chin. J. Appl. Ecol. 26(8), 2371–2378 (2015)Google Scholar