On-line Inspection Sugar Content and Acidity in Fruits Using Near Infrared Technology

• Main Authors: Chia-Tseng Chen, 陳加增
• Other Authors: Suming Chen
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/23608711788698862285

碩士 === 國立臺灣大學 === 生物產業機電工程學研究所 === 89 === This study focuses on using near infrared (NIR) technology to investigate the sugar content and acidity in grape and mango using pre- and post-dispersive spectrophotometers. The calibration models of sugar content and acidity are developed and applied to design an on-line NIR inspection system. The spectra of juices and intact fruits were scanned and analyzed, and then the results points out that all the best analyzed model were modify partial least square regression (MPLSR) instead of the best grape juice transmittance spectra absorption of sugar content is multiple linear regression (MLR) with five wavelengths combination (2272, 2280, 1874, 1732, and 1436 nm) in the second derivative, which is rc=0.991, SEC=0.195, rp=0.990, and SEP=0.210. The best result of in grape juice is given by MPLSR method (first derivative , wavelength range: 800~1000＋1300~1500＋1600~1900＋2350~2450nm), which is rc=0.982, SEC=0.023, rp=0.976, and SEP=0.026. The post-dispersive model gets better results than the pre-dispersive in the analysis of pulp reflection spectra, and the results about sugar content and acidity of graph is better than mango. The best results of sugar content of graph are rc=0.961, SEC=0.416, rp=0.950, and SEP=0.463 (second derivative, wavelength range: 800~1100 nm), and the best acidity result is rc=0.935, SEC=0.042, rp=0.894, and SEP=0.052 (first derivative, wavelength range: 400~2500 nm). The best result of sugar content of mango is rc=0.938, SEC=0.601, rp=0.915, and SEP=0.649 (first derivative, wavelength range: 700~1300 nm), and the acidity result is rc=0.782, SEC=0.031, rp=0.749, and SEP=0.030 (first derivative, wavelength range: 500~2100 nm). The developed on-line NIR inspection system for measuring sugar content and acidity in fruits, which using post-dispersive model and dynamic-data-exchange (DDE) programming, was successfully designed. This computer controlled integrates NIR scanning, conveying mechanism, program logical controller and computer infacing. The investigation of system performance gave satisfactory results.