近红外分析花生籽粒脂肪酸含量的研究

RESEARCH ON FATTY ACID CONTENT IN PEANUT SEED BY NEARINFRARED REFLECTANCE SPECTROSCOPY

采集了50种具有代表性的花生品种作为样本,建立了花生籽粒5种主要脂肪酸(棕榈酸、油酸、亚油酸、花生酸和山嵛酸)含量的近红外光谱定标模型.其中,油酸和亚油酸的定标模型质量较高,交互验证决定系数分别为0.933 0和0.924 9,外部验证决定系数分别为0.940 1和0.948 7,相对分析误差均大于2.5.棕榈酸、花生酸和山嵛酸的定标模型交叉验证决定系数和外部验证决定系数均小于0.8,相对分析误差均小于2.5.试验结果表明应用近红外光谱法测定花生籽粒中油酸和亚油酸这两种脂肪酸的含量模型预测的精确度较高.

Fifty representative peanut cultivars were selected as samples to construct a near-infrared reflectance spectroscopy （NIRS） calibration model for predicting the contents of five main fatty acids in the peanut seeds, such as palmitic acid, oleic acid, linoleic acid, behenic acid and arachidic acid. The oleic acid and linoleic acid calibration models had higher quality; the cross validation decision coefficients were respectively 0.933 0 and 0.924 9; the external validation decision coefficients were respectively 0.940 1 and 0.948 7; and the relative percent deviations of the both two models were larger than 2.5. The cross validation decision coefficients and extemal validation decision coefficients of the calibration models for palmitic acid, behenic acid and arachidic acid were less than 0.8; and the relative percent deviations were all less than 2.5. The test results showed that the models had high precision in predicting oleic acid and linoleic acid contents determined by NIRS.