夏橙的含糖量是决定夏橙口感的重要因素。通过试验研究,采用近红外漫反射光谱分析技术实现对夏橙糖度间接测定的可行性。首先通过对夏橙不同放置状态下进行光谱检测,并建模分析,发现当夏橙侧躺采集光谱的测定精度和稳定性最优,但采用直立放置和任意放置的结果也能满足实际需求,说明夏橙的放置状态不会对近红外测定结果造成太大的影响;然后通过对样品进行分组,分别在不用温度下采集夏橙的漫反射光谱,建立夏橙糖度的预测模型,并且进行组内和组间交叉验证,结果发现,当温度变化幅度不是很大时,夏橙糖度的预测模型具有良好的精度,符合实际工作环境的要求。即采用近红外漫反射技术实现对夏橙糖度的无损在线检测是可行的。
The effects of different place state (standing, side-lying, and being placed in arbitrary) on sugar concentration of Valencia orange were studied. The results of side-lying were slightly better than those of the other two modes. Results of the other place states can meet practical requirement by near infrared diffuse reflectance. The grouping of Valencia orange samples was collected by diffuse reflection spectra based on different temperatures. The prediction model of sugar concentration was established on cross validation between group and within group. When the temperature variation was trivial, the prediction model of Valencia orange sugar concentration had good precision, fitting well with the requirements of actual work environment.