The goal of this research was to develop a portable system that could be used to evaluate the quality of milk in real time at a raw milk production site. A real-time portable quality evaluation system for raw milk was developed to enable non-destructive quality evaluation of somatic cell count (SCC), fat, protein, lactose, and total solid (TS) in milk samples. A prediction model of SCC, fat, protein, lactose, and TS was constructed using partial least squares (PLS) and 200 milk samples were used to evaluate the prediction performance of the portable quality evaluation system and high performance spectroscopy. Through prediction model development and verification, it was found that the accuracy of high performance spectroscopy was 90% for SSC, 96% for fat, 96% for protein, 91% for lactose, and 97% for TS. In comparison, the accuracy of the portable quality evaluation system was relatively low, at 90% for SSC, 95% for fat, 92% for protein, 89% for lactose, 92% for TS. However, the measurement time for high performance spectroscopy was 10 minutes for 1 sample, while for the portable quality evaluation system it was 6 minutes. This means that the high performance spectroscopy system can measure 48 samples per day (8 hours), while the portable quality evaluation system can measure 80 (8 hours). Therefore, it was found that the portable qual ity evaluation system enables quick on-site quality evaluation of milk samples.

This study was conducted to develop a portable quality evaluation system of bee-honey by near infrared spectroscopic technique. Two kinds of bee-honeys from acacia and polyflower sources were tested in this study. The system consists of power supply, tungsten-halogen lamp, detector, and optical fiber probe. Performance of the system was analyzed by comparing the prediction accuracy of the laboratory spectrophotometer. Total of 346 spectra was divided into a calibration set and a prediction set. The PLS (Partial Least Squares) models were developed to predict the quality parameters of bee-honeys. Reflectance spectra, moisture contents, ash, invert sugar, sucrose, F/G ratio, HMF(hydroxy methyl furfural), and ratio of honeys were measured. The PLS models of the laboratory spectrophotometer showed good relationships between predicted and measured quality parameters of honeys in the wavelength range of 1.100~2.200 nm. The PLS analysis of the portable quality evaluation system showed good relationships between predicted and measured quality parameters of honeys in the wavelength range of 1.100~1.300 nm and 1.400~1.700 nm. The results showed the feasibility of the portable quality evaluation system to determine the quality parameters of bee-honey in the field during harvesting.