Abstract:

This study developed a nondestructive test method based on laser backscattering image to monitor the food value
of ‘Xiacui’ peach during shelf life. Classification criteria of food value for ‘Xiacui’ peach were established based on sensory
score. By detecting the changes in hardness, soluble solids content (SSC) and titratable acid for each peach, this classification
method was proved to be credible. Moreover, a semiconductor laser generator with 650 nm wavelength and 20 mW power
was used to build laser speckle measurement device. Laser backscattering images were captured and analyzed to extractcolor
(R, G, B, H, S, I and each standard deviation) and texture parameters (pixel area, average gray value, standard deviation,
smoothness, third-order moments, consistency and entropy). The changes in pixel area, R, G, B, consistency and entropy
were correlated with the food value of ‘Xiacui’ peach. Finally, these six characteristic parameters were chosen as input to
develop a support vector classification (SVC) model with overall training and testing classification accuracy of 95.0% and
92.5%, respectively. Our findings suggest that laser backscattering imaging can be useful to non-destructively detect the food
value of ‘Xiacui’ peach during shelf life.

Key words: laser backscattering imaging, peach, food value, monitoring