基于计算机视觉食品中大肠杆菌快速定量检测_殷涌光.pdfVIP

39 2 吉林 大学 学 报 ( 工学 版) V ol. 39 Sup. 2 2009 9 J ournal o f Jilin U niv ersity ( Engineering and T echnolo gy Edition) S ept . 2009 殷涌光, 丁 筠 ( 吉林大学 生物与农业工程学院, 长春 130022) 利用大肠杆菌能发酵乳糖产酸从而使伊红 兰混合液产生沉淀溶液变色的特性, 设 计了一套基于颜色特征识别技术的食品中大肠杆菌快速定量检测系统, 通过16 h 培养后溶液 颜色的变化程度来判断待测液中大肠杆菌的数目系统自动提取发酵后溶液图像的H IS 颜色分量作为输入向量, 自动调用训练好的BP 神经网络模型得到大肠杆菌数试验结果表 明该方法的检测结果与传统方法的相关性好, 检测时间在 18 h 以内, 远少于传统方法的6 天, 有效提高了产品的销售品质 食品机械; 计算机视觉; 大肠杆菌快速检测; 人工神经网络; 颜色特征识别 T S207. 4; T P242 . 6 A 1671- 5497( 2009) Sup. 2- 0 344- 05 Rapid methodfor enumeration of Escherichia coli in food based on computer vision Y IN Y ong- gu ang , DIN G Y un ( College of Biological and A gricultural Engineering , Jilin University , Changchun 130022, China) Abstract:T he sp ecialit y th at the Escherichia co li can f erment lact ose to pro du ce acid w hich can pro du ce depo sit io ns and make t he indicat or s co lor ch anged w as ut ilized, an aut om atic det ect ion syst em fo r enumeration of Escherichia co li in f ood based on color feature recog nit ion t echnolo gy w as designed and the color change deg ree af t er incubat ing f or 16 h w as used to ju dg e t he Escher ichia co li count s in the det ectio n samples. T he syst em can aut om at ically ex t ract t he H IS colo r vect or t o act as t he input vect or and t ran sf er the t rained BP neural net w ork t o g et t he Escherichia co li counts . T he ex periments indicate th at comp ar ison s of co unt s of Escherichia coli by t he rapid aut omat ic detection system and by the t radit ional met ho d closed correlat ed . M oreov er, by u sing t his rapid