农产品加工(下半月)
農產品加工(下半月)
농산품가공(하반월)
AEM RODUCTS ROCESSING
2015年
1期
47-49,53
,共4页
陈醋%可溶性固形物%近红外光谱%偏最小二乘法
陳醋%可溶性固形物%近紅外光譜%偏最小二乘法
진작%가용성고형물%근홍외광보%편최소이승법
aged vinegar%soluble solids%NIRS%PLS
可溶性固形物作为陈醋的重要检测指标,可以反映醋品质的优劣。选择宁化府、水塔、东湖、紫林4个品牌各60瓶老陈醋作为研究对象,选用ASD公司的FieldSpec 3型光谱仪直接对这240瓶瓶装陈醋进行透射检测,分别采用偏最小二乘法(PLS)在800~2000 nm波段内对Splice Correction校正后的光谱数据、一阶导数和二阶导数光谱数据建立了可溶性固形物定量预测模型。结果表明,经Splice Correction校正后的光谱数据建模效果最好,校正集样本的相关系数为0.86,标准偏差为0.6217,标准差为0.6288;预测集样本的相关系数为0.91,标准偏差为0.7530,预测标准差为0.7462。说明本试验所建模型可以预测瓶装陈醋中的可溶性固形物含量,采用近红外光谱结合PLS分析方法快速测定瓶装陈醋中可溶性固形物的方法是可行的。
可溶性固形物作為陳醋的重要檢測指標,可以反映醋品質的優劣。選擇寧化府、水塔、東湖、紫林4箇品牌各60瓶老陳醋作為研究對象,選用ASD公司的FieldSpec 3型光譜儀直接對這240瓶瓶裝陳醋進行透射檢測,分彆採用偏最小二乘法(PLS)在800~2000 nm波段內對Splice Correction校正後的光譜數據、一階導數和二階導數光譜數據建立瞭可溶性固形物定量預測模型。結果錶明,經Splice Correction校正後的光譜數據建模效果最好,校正集樣本的相關繫數為0.86,標準偏差為0.6217,標準差為0.6288;預測集樣本的相關繫數為0.91,標準偏差為0.7530,預測標準差為0.7462。說明本試驗所建模型可以預測瓶裝陳醋中的可溶性固形物含量,採用近紅外光譜結閤PLS分析方法快速測定瓶裝陳醋中可溶性固形物的方法是可行的。
가용성고형물작위진작적중요검측지표,가이반영작품질적우렬。선택저화부、수탑、동호、자림4개품패각60병로진작작위연구대상,선용ASD공사적FieldSpec 3형광보의직접대저240병병장진작진행투사검측,분별채용편최소이승법(PLS)재800~2000 nm파단내대Splice Correction교정후적광보수거、일계도수화이계도수광보수거건립료가용성고형물정량예측모형。결과표명,경Splice Correction교정후적광보수거건모효과최호,교정집양본적상관계수위0.86,표준편차위0.6217,표준차위0.6288;예측집양본적상관계수위0.91,표준편차위0.7530,예측표준차위0.7462。설명본시험소건모형가이예측병장진작중적가용성고형물함량,채용근홍외광보결합PLS분석방법쾌속측정병장진작중가용성고형물적방법시가행적。
As an important detection index to aged vinegar, soluble solids can reflect aged vinegar's quality. Four brands of aged vinegar, such as Ninghuafu, Shuita, Donghu and Zilin are chosen as the research objects, 60 bottles are included in each brand. Portable optical emission spectrometer FieldSpec 3 produced by ASD is used to transmit and detect these 240 bottled vinegar samples, and Splice Correction as spectrum preprocessing method is applied to eliminate the interference during the spectra diction. Quantitative prediction model based on soluble solids is then built by using Partial Least Square (PLS) regression methods under 800~2 000 nm. The results show that the correlation coefficient of modeling samples in bottled aged vinegar is 0.86, the prediction samples is 0.91, the standard deviation of calibration samples is 0.621 7, the prediction samples is 0.753 0, the standard deviation of calibration is 0.628 8, the prediction is 0.746 2. The model built in this experiment can predict the soluble solids content in bottled vinegar and indicate that it is feasible to predict soluble solids of bottled aged vinegar by using NIRS and PLS.