Abstract:

Hyperspectral half-transmittance imaging technique was investigated for the detection of chilling injury in ‘Xiahui
5’ peach during storage. Sensory score, browning index, hardness and extractable juice content were measured. At the same
time, classification standard of chilling injury for ‘Xiahui 5’ peach was established based on the correlation of sensory score
with physicochemical values. Then, hyperspectral half-transmission images over the wavelength range of 400–1 000 nm
were captured. Independent component analysis (ICA) and weighting coefficients for ICA were implemented on the
pictures. The optimal wavelengths selected were 640, 745, and 811 nm. Hyperspectral half-transmission ICA images for
different injury levels of chilled peach were obtained, and black spots were found which indicated chilling injury. Finally,
the spectral average of each characteristic band was chosen as the input of the Fisher discrimination model, and an average
testing accuracy of 91.0% was achieved to distinguish chilled ‘Xiahui 5’ peach with different injury levels. This research
demonstrated that the hyperspectral half-transmittance imaging technique is feasible for the detection of chilled peach at
different injury levels.

Key words: hyperspectral imaging, half-transmittance, peach, chilling injury, detection