Abstract Detecting mature-green and immature-green tomatoes using non-destructive approaches is a challenge for the fresh produce industry. Hyperspectral fluorescence imaging technique with excitation wavelength at 365 nm and UV–vis CCD camera was used for early non-destructive detection of mature-green and immature-green fruit from 200 randomly harvested green tomatoes. Conventional destructive analysis regarding locule gel development and seed texture were assessed to assign the maturity stage of the fruit. In addition soluble solid content (SSC), pH, total acidity (TA), and color were measured, on the training set and on the prediction set, in this case also after 10 d of storage. Fluorescence intensity at the surface of immature-green fruit was higher in the red region (690 nm) than that of mature-green fruit, suggesting that hyperspectral fluorescence imaging can be an effective classification tool. A univariate classification method was used to distinguish mature-green and immature-green tomatoes based on the grey scale values extracted from fluorescence imaging, with a non-error rate of 96 % in calibration and 100 % in external prediction. Hence, a non-destructive method for the early distinction of mature-green from immature-green tomatoes is available.

中文翻译：

---

使用荧光成像方法早期区分成熟和未成熟的绿色番茄（Solanum lycopersicum L.）

摘要 使用无损方法检测成熟绿色和未成熟绿色西红柿是新鲜农产品行业面临的挑战。使用激发波长为 365 nm 的高光谱荧光成像技术和 UV-vis CCD 相机对来自 200 个随机收获的绿色西红柿的成熟绿色和未成熟绿色水果进行早期无损检测。对关于室凝胶发育和种子质地的常规破坏性分析进行了评估，以指定果实的成熟阶段。此外，还测量了训练集和预测集的可溶性固形物含量 (SSC)、pH、总酸度 (TA) 和颜色，在这种情况下也是在储存 10 天后。未成熟绿色果实表面的荧光强度在红色区域（690 nm）高于成熟绿色果实，表明高光谱荧光成像可以成为一种有效的分类工具。基于荧光成像提取的灰度值，采用单变量分类方法区分成熟绿色和未成熟绿色番茄，校准无错误率为96%，外部预测无错误率为100%。因此，有一种用于早期区分成熟绿色和未成熟绿色番茄的非破坏性方法是可用的。