利用近紅外光技術檢測稻米鮮度之研究

Abstract

稻米的貯藏時間增加對於稻米之外觀、食味以及營養品質影響甚鉅，故稻米之鮮度檢測為農委會於現場檢測之重點項目之一。近紅外光光譜分析技術因具有非破壞性及快速檢測之優點，已廣泛地被應用於各科學領域上。研究中分別使用NIRS 6500型分光光度計搭配自動傳送樣本配件和NIRS 6500型分光光度計搭配RCA配件量取台南11號之稻穀、糙米與白米的近紅外光光譜資訊，並搭配SIMCA及PLS-DA兩種定性分析方法判別不同鮮度之稻米樣本。並利用現行檢測稻米鮮度BTB-MR多粒米試管法檢測白米酸鹼值，以及美國穀物化學協會所制定之方法檢測稻米脂肪酸度含量，並以MPLSR及MLR方法建立白米酸鹼值與脂肪酸度之檢量模式。本研究建立白米酸鹼值與脂肪酸度之轉換公式，可利用白米酸鹼值轉換得出脂肪酸度含量。 分類不同貯藏期別稻米之研究顯示，稻穀之成功率皆為100 %；糙米以PLS-DA分類之最佳平均成功率為95 %，而SIMCA分類最佳成功率則為90 %；白米以PLS-DA分類之最佳平均成功率為86.7 %，SIMCA分類之最佳成功率則為88.3 %。另以稻米酸鹼值大小訂出不同鮮度之稻米組別分類結果中，糙米以PLS-DA分類之最佳成功率為92.33 %，SIMCA分類最佳成功率則為100 %；白米以PLS-DA分類之最佳成功率為93 %，SIMCA分類之最佳成功率則為100 %，整體分類成功率均優於以不同收穫期別分組之結果。最後，再以脂肪酸度含量所區分的三組不同鮮度稻米，也可以PLS-DA分析法100 %成功分類不同鮮度的糙米與白米。以上結果顯示欲以定性分析法進行不同鮮度稻米之分類，應以稻米酸鹼值或脂肪酸度含量作為判定指標為佳。 在定量分析中，與白米酸鹼值進行MPLSR分析之結果，最佳檢量線rc皆達0.991，SEC皆為0.071；而與白米酸鹼值進行MLR分析之結果，以搭配RCA配件量測之結果，以一次微分光譜並選擇8個波長數最佳，rc可達0.992，SEC為0.067，顯示兩種檢量模式皆具有相當不錯之預測能力。白米脂肪酸度之定量分析結果，則以自動傳送樣本配件搭配MPLSR分析時，rc可達0.986，SEC為0.116，顯示以定量分析方式預測白米脂肪酸度含量之可行性很高。 利用近紅外光檢測技術，可成功分類不同鮮度之稻米樣本，且可準確預測其白米酸鹼值與脂肪酸度含量，另外也證實白米酸鹼值與脂肪酸度比稻米的年份期別更適合作為評量稻米鮮度的重要參考指標，並可確實達到檢測稻米鮮度之目的。相對於傳統慣用之目視比色法判斷稻米酸鹼值，本研究所建立之近紅外光檢驗法可應用於現場之量測，為一種快速、方便、非破壞性且準確的方法。

The storage time of rice has an enormous affect on its appearance, flavor, and quality of the nutrients. Therefore, the determination on freshness of rice is one of the main goals of Council of Agricultural in site examination. Near-infrared spectroscopy has been extensively applied in many scientific fields, due to its advantages of non-destructive and rapid measurements. This research aims to study the rice freshness in terms of qualitative and quantitative approaches using near-infrared spectroscopy. The rice samples include paddy rice, brown rice and milled rice of TN-11 variety. To distinguish levels of rice freshness, SIMCA and PLS-DA were used in qualitative approach while MPLSR and MLR were adopted in quantitative approach. The rice freshness is expressed by pH value and fat acidity of rice in this study. The values of pH value are determined by BTB-MR method, and fat acidity by the rapid method established by American Association of Cereal Chemists. The formula to convert pH value into fat acidity was established in this study. Rice freshness in terms of different harvest period and year, the successful identification rate (SIR) of paddy rice is 100%. According to the results of brown rice, the SIR of PLS-DA classification is 95% while the SIMCA classification is 90%. Among the milled rice samples, the SIR of PLS-DA classification is 86.7% while the SIMCA classification is 88.3%. Regarding the results of using pH value to represent rice freshness, the SIR of PLS-DA classification is 92.33% and the SIMCA classification is 100% for brown rice. As for milled rice, the SIR of PLS-DA classification is 93% and the SIMCA classification is 100%. In general, the values of SIR in terms of pH value are better than those by harvest year. Freshness distinguished by the content of fat acidity, SIR can reach 100 % for both brown and milled rice.It can be concluded that both pH value and fat acidity be used as criteria to represent the freshness of rice In quantitative approach based on the determination of pH value, the best MPLSR model has reached correlation coefficient rc of 0.991 and SEC of 0.071;the best MLR model reveals the results of rc = 0.992 and SEC =0.067. The best MPLSR model based on the determination of fat acidity yields the results of rc = 0.986 and SEC =0.116. With near-infrared spectroscopy, rice freshness can be successfully classified; the pH value and fat acidity of milled rice can be predicted quantitatively. It is also confirmed that pH value and fat acidity are more appropriate to serve as indicators of rice freshness than harvest time. On contrary to the traditional visual colorimetric inspection of the pH value, near-infrared models developed in this study is a good method with rapid, convenient, non-destructive and precise fearures in real fields applications.