Over 40,000 discrete samples of soybeans (Glycine max) were obtained from locations around the United States grown during the years 2013–2018. The samples were ground and scanned on near infrared spectrometer (NIRS). The carbohydrate composition—sucrose, raffinose family oligosaccharides, total soluble sugar composition, total insoluble sugars, and total uronic acids—was determined by prediction using NIRS calibration models. Multiple linear regression models were developed using about 2,300 soybean samples and an additional 700 samples for validation with NIRS scans and carbohydrate composition measurements using analytical methods. Relationship between carbohydrates and the agronomic traits of the soybeans were determined. It was found that crop year, maturity group, and location had a significant effect on the carbohydrate composition of soybeans. NIRS is an accurate, rapid, and nondestructive method to determine carbohydrate composition in large numbers of soybeans.

中文翻译：

---

近红外光谱在确定大豆作物年份、成熟组、位置和碳水化合物组成之间的关系中的应用

超过 40,000 个离散的大豆样品（最大甘氨酸) 来自美国各地在 2013-2018 年间种植的地方。将样品研磨并在近红外光谱仪 (NIRS) 上扫描。碳水化合物组成——蔗糖、棉子糖家族寡糖、总可溶性糖组成、总不溶性糖和总糖醛酸——通过使用 NIRS 校准模型的预测来确定。使用大约 2,300 个大豆样品和另外 700 个样品开发了多元线性回归模型，用于使用分析方法通过 NIRS 扫描和碳水化合物成分测量进行验证。确定了碳水化合物与大豆农艺性状之间的关系。结果表明，作物年份、成熟组和地点对大豆的碳水化合物组成有显着影响。NIRS 是一种准确、快速、