Nondestructive detection of sugar content in early season maturity pear based on half transmittance technique

A portable fiber spectrometer was used for constructing a spectra detecting system based on half transmittance technique. The half transmission spectral system was applied for measurement sugar content in early season maturity pear. Different spectral pre-processing methods of first derivative [D(%T)], second derivative [D²(%T)], and smoothing were used for establishing three different models based on multiple linear regression method(MLR), principle component regression method(PCR), and partial least square method (PLS). The results showed that the model based on PLS after second derivative [D² (%T)] pre-processing obtained an appropriate performance. The model with respect to PLS was able to predict the sugar content in early season maturity pear with good correlation(R²=0.72), and standard error of prediction(RSEP= 0.41). The result showed it is a portable way to detect sugar content in early season maturity pear based on half transmittance technique.