於麵粉生產過程進行取樣後,先以不同篩目(40、80、100、1 20、14及200目)之篩網篩選後,再以近紅外線光譜儀進行水分、灰分、白度、筋度及蛋白質等品質分析。結果得知以近紅外線光譜儀進行不同粒度大小之麵粉掃瞄時,樣品粒度越小,其光譜再現性越佳。另外,以不同粒度大小之麵粉所建立近紅外線光譜儀品質分析,包括水分、灰分、白度、筋度及蛋白質之積量線,發現以140及200目之樣品有最佳的預測結果,其預測標準偏差值(Standard Error of Performance,SEP)分別低於0.13,0.03,0.33,0.27及0.1,其測定值與傳統分析值之相關係數( Correlation coefficient,r)均高於0.95。
Sampling flour with different particle size by through different mesh No. 40, 80, 100, 120, 140 and 200 during processing. The quality analyses including water content, ash, white degree, gluten degree and protein content by using Near Infrared Spectroscopy(NIRS) Technique. Results show the small in particle size of flour, the less interference in NIRS scanning. Calibration curves for quality analyses of flour by NIRS were established. Results show that their R^2 values are all higher than 0.98 except for water con- tent was 0.92. The correlation coefficients(r values) of their predicted and measured values are all higher than 0.95 for samples with particle size 25- 60 and 35-75 um(200 and 140 mesh) and the standard error of performance(SEP) were less than 0.13, 0.03, 0.33, 0.27 and 0.1 for water content, ash, white degree, gluten degree and protein content, respectively.