One of the major compounds in almond kernels, which determines their nutritional quality, are lipids. The aim of this research was to determine the fatty acid profile in intact in-shell and shelled almonds (145 samples) using two new generation handheld near infrared spectroscopy (NIRS) sensors, of different optical design and technical specifications, adapted for in situ analysis in different stages in the food supply chain: in the industry after harvest, at the reception points and during postharvest storage. For both instruments, two procedures for taking near infrared (NIR) spectra were tested: (1) static, where point spectral readings were taken of almonds placed on trays; (2) dynamic, where spectra were taken by scanning the entire trays. Modified partial least squares (MPLS) regression models were developed using NIR spectra with different combinations of signal pre-treatments — derivative and scatter correction methods. The residual predictive deviation for cross validation (RPD_cv) of the best models developed for the prediction of palmitic, stearic, oleic, and linoleic acids using shelled almonds were 2.40, 2.16, 3.98, and 3.77, respectively, and 1.73, 1.73, 2.02, and 2.11 for the in-shell almonds. These results confirm the feasibility of NIRS technology to measure the fatty acid profile in in-shell and shelled almonds. A comparison between the presentation mode (in-shell or shelled) and analysis mode (static or dynamic) showed that the best results were obtained for shelled almonds analysed in dynamic mode.

脂质是杏仁仁中决定其营养品质的主要化合物之一。这项研究的目的是使用两个适用于原位的不同光学设计和技术规格的新一代手持式近红外光谱（NIRS）传感器，确定完整的带壳和带壳杏仁（145个样品）中的脂肪酸谱食品供应链中不同阶段的分析：收获后，接收点和收获后储存的行业。对于这两种仪器，都测试了两种获取近红外（NIR）光谱的程序：（1）静态，对放在托盘上的杏仁进行点光谱读数；（2）动态的，其中通过扫描整个托盘获取光谱。使用NIR光谱和信号预处理的不同组合（导数和散点校正方法）开发了改进的偏最小二乘（MPLS）回归模型。交叉验证的剩余预测偏差（RPD _cv）使用去壳杏仁预测棕榈酸，硬脂酸，油酸和亚油酸的最佳模型分别为2.40、2.16、3.98和3.77，而带壳杏仁的预测分别为1.73、1.73、2.02和2.11。这些结果证实了NIRS技术用于测量带壳和带壳杏仁中脂肪酸谱的可行性。展示模式（带壳或带壳）和分析模式（静态或动态）之间的比较表明，以动态模式分析的带壳杏仁获得了最佳结果。