Almond Board of California , 1150 Ninth Street, Suite 1500, Modesto, California 95354, United States.
J Agric Food Chem. 2016 Jul 27;64(29):5958-62. doi: 10.1021/acs.jafc.6b01828. Epub 2016 Jul 18.
Concealed damage (CD) is defined as a brown discoloration of the kernel interior (nutmeat) that appears only after moderate to high heat treatment (e.g., blanching, drying, roasting, etc.). Raw almonds with CD have no visible defects before heat treatment. Currently, there are no screening methods available for detecting CD in raw almonds. Herein, the feasibility of using near-infrared (NIR) spectroscopy between 1125 and 2153 nm for the detection of CD in almonds is demonstrated. Almond kernels with CD have less NIR absorbance in the region related with oil, protein, and carbohydrates. With the use of partial least squares discriminant analysis (PLS-DA) and selection of specific wavelengths, three classification models were developed. The calibration models have false-positive and false-negative error rates ranging between 12.4 and 16.1% and between 10.6 and 17.2%, respectively. The percent error rates ranged between 8.2 and 9.2%. Second-derivative preprocessing of the selected wavelength resulted in the most robust predictive model.
隐性损伤（CD）定义为核仁内部（果仁）的棕色变色，仅在中度至高度热处理（如烫漂、干燥、烘烤等）后出现。具有 CD 的生杏仁在热处理前没有可见的缺陷。目前，尚无用于检测生杏仁中 CD 的筛选方法。本文证明了在 1125 至 2153nm 近红外（NIR）光谱范围内检测杏仁中 CD 的可行性。具有 CD 的杏仁核在与油、蛋白质和碳水化合物相关的区域具有较低的 NIR 吸收率。使用偏最小二乘判别分析（PLS-DA）和选择特定波长，开发了三个分类模型。校准模型的假阳性和假阴性错误率分别在 12.4%至 16.1%和 10.6%至 17.2%之间。误差率在 8.2%至 9.2%之间。所选波长的二阶导数预处理产生了最稳健的预测模型。
