摘要：Anthocyanin, one of the major secondary metabolites in red wave lettuce, is beneficial for human health and improvement of visual functions in salad. While the anthocyanin measurement in plant mainly relies on chemical analysis to date. In order to explore the possibility of digital image in non-destructive anthocyanin prediction, a quadratic model based on color parameters (red-green-blue color space and hue-saturation-intensity color space, RGB and HSI color space) of lettuce leaf was investigated. The results suggested that the characteristics of different color components combinations such as R/G, G/R, B/G, G/(R + B), G/(R + G + B), H, I/H, S/H, H/S, R/(R + G + B) - G/(R + G + B), G/(R + G + B) - B/(R + G + B), (G - R) /(G + R) and (G - B) /(G + B) were significantly correlated with anthocyanin content, and the correlation coefficient between S/H and anthocyanin content was the highest of 0.850. For quantitative prediction of anthocyanin content, B/G value was optimal color characteristic to forecast the anthocyanin content, with the function of Y=aX+bX2+c (R2 of 0.781). This work demonstrates that the machine vision technique is promising non-destructive measurement of anthocyanin content in lettuce leaves.
