CORRELATION BETWEEN GLUCOSINOLATE CONTENT AND SPECTRAL REFLECTANCE OF CABBAGE LEAVES USING A SPECTROMETER

Reflectance properties of cabbage leaves at ultra-violet/visible/near-infrared wavebands were studied to examine the relationship between reflectance at selected wavelengths and glucosinolate content for remote sensing.
A total of 63 cabbage leaves were sampled from a plant factory to collect reflectance spectra with an UV/VIS/NIR spectrometer, and then the glucosinolate contents were determined through labora¬tory analysis.
The collected reflectance spectra in a range within 250-1100 nm were used to analyze relationship between the reflectance spectra and glucosinolate content.
The spectra were collected using the software provided by the manufacturer.
To reduce spectral noise, Savitzky-Golay smoothing method with 15 factors on a subset of the dependent variables was used to preprocess the reflectance data.
Before the smoothing method was applied, mathematical transformation, including exponential reflectance (Exp(R)), the square of reflectance (R²), the square root of reflectance (√R), and inverse reflectance (1/R) was applied.
Then, stepwise multiple linear regression (SMLR) models were obtained.
Good relationships were obtained for neoglucobrassicin, glucoraphanin, and 4-methoxyglucobrassicin, while a poor relationship was obtained for the total glucosinolate content.
Important wavelengths selected were 305.46 nm, 310.26 nm, 357.21 nm, 414.4 nm, 446.58 nm, 451.21 nm, 485.87 nm, 1,013.3 nm, and 1,027.57 nm.

International Symposium on New Technologies for Environment Control, Energy-Saving and Crop Production in Greenhouse and Plant Factory – Greensys 2013

V.D. Ngo, D.K. Ryu, S.W. Kang, S.O. Chung, S.U. Park, S.J. Kim, J.T. Park

wavelength selection, optical reflectance, mathematical transformation, correlation, SMLR

https://doi.org/10.17660/ActaHortic.2014.1037.33