研究目的
To evaluate the ability of visible to near infrared spectroscopy (Vis/NIRS) to non-destructively quantify total soluble solids (TSS), titratable acidity (TA), TSS:TA ratio, and BrimA of 'Valencia' orange and 'Star Ruby' grapefruit, and to demonstrate its potential for commercial application in predicting internal quality parameters.
研究成果
Vis/NIRS is a fast and accurate technique for non-destructively predicting internal quality parameters like BrimA, TSS, TA, and TSS:TA ratio in citrus fruit. It can be used for fruit discrimination during packing and pricing, benefiting producers and consumers by enabling pricing based on taste rather than physical attributes.
研究不足
The study noted that fruit variability within the same batch may cause bias in conclusions based on destructive measurements, and prediction accuracy was lower for grapefruit compared to oranges. The models may require further validation for broader application.
1：Experimental Design and Method Selection:
The study used Vis/NIRS combined with chemometrics to develop prediction models for internal quality parameters. Spectra were collected in reflectance mode, and chemometric analysis included principal component analysis (PCA) and partial least squares (PLS) regression with test set validation.
2：Sample Selection and Data Sources:
A total of 120 'Star Ruby' grapefruit and 120 'Valencia' oranges were randomly picked from two orchards in South Africa, mixing all possible sizes of mature fruit.
3：List of Experimental Equipment and Materials:
Equipment included a NIRSystems Model XDS spectrometer, a digital refractometer (RFM340+), a stirrer (ULTRA-TURRAX IKA T25 digital), and chemicals like sodium hydroxide. Materials included fruit samples, distilled water, and plastic specimen jars.
4：Experimental Procedures and Operational Workflow:
Fruit were equilibrated at room temperature for 24 hours. Vis/NIR spectra were acquired from two opposite sides of each fruit. After scanning, fruit were peeled, juice was collected and homogenized, TSS was measured with a refractometer, and TA was measured by titration. BrimA and TSS:TA ratio were calculated.
5：Data Analysis Methods:
Statistical analysis used GenStat software. Chemometric analysis used Vision software for PCA and PLS regression, with spectral preprocessing including standard normal variate (SNV) and Savitzky-Golay smoothing. Model performance was assessed using R2, RMSEP, RPD, and bias.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
