NON-DESTRUCTIVE BRIX SENSING OF FLORIDA GRAPEFRUITAND HONEY TANGERINE

摘要

Preliminary tests were undertaken to assess the accuracy of near infra-red (NIR) technologies to measure Brix level of intact Florida grapefruit. Two units were utilized: (1) an online prototype unit (Mitsui QScope, Sumitomo Metal Mining Co., Tokyo, Japan) and (2) an hand-held unit developed for general fresh fruit quality assessment (Inst. of Agric. Engr.-Bornim, Potsdam, Germany). The on-line tests were conducted at a rate of ~5 fruit per second. Pre-conditioned Interior white grapefruit were tested at two temperatures, 5 and 20 °C while commercial Indian River red grapefruit were tested at ambient conditions. Comparative measurements were obtained for both on-line and hand-held NIR units with the fruit subsequently juiced to obtain a refractiveindex measurement of the fruit's Brix level. Linear regression relationships were developed between the non-destructive NIR techniques and the laboratory Brix measurements. The highest correlation, r~2 = 0.67**, was found for Indian River red grapefruitof the smaller size of fruit tested (9.5-11.0 cm diameter). Percent correct classification analysis was ascertained at 9 °Brix for Interior grapefruit and 10 °Brix for Indian River red grapefruit. Higher classification rates were obtained for ambient temperature tests at -20 °C and for the smaller size of Indian River red grapefruit. A neural network (NN) model with combined inputs of physical and color attributes, plus predicted Brix, was evaluated. Corresponding Brix data for Honey tangerine also were obtained utilizing a handheld NIR probe.