蔬菜中大肠杆菌的机器视觉快速检测

Rapid Detection Based on Machine Vision for Escherichia coli in Vegetables

为了适应蔬菜等农产品对大肠杆菌快速检测的需求,提出采用形态特征参数及染色后菌体区域的颜色特征参数统计值对大肠杆菌进行快速识别,同时提出采用主成分神经网络作为预测模型来提高识别能力。提取了Hu’s不变矩、形状因子、密集度、饱和度等14个具有尺度、平移、旋转不变性的特征参数,提取主成分建立了基于主成分的3层BP神经网络模型。将其与普通神经网络模型比较的结果表明,主成分神经网络简化了网络结构、减少了训练时间和计算量、提高了识别的正确率,对大肠杆菌的识别正确率达到91.33%。

In order to adapt to the requests of on-site rapid detection technique of Escherichia coli（E.coli） for the safety of agricultural products,a rapid E.coli recognition method based on shape and color feature parameters was proposed.Principal component neural network was used to improve the recognition ability.Principal component analysis was applied to the 14 extracted feature parameters,including Hu＇s moment invariants,shape factor,denseness and saturation,et al.A three-layer BP neural network model based on the principal components was constructed.Compared with traditional BP neural network,the configuration of the principal component neural network was simpler,the training time was shorter and the recognition accuracy was higher.The recognition accuracy of the principal component neural network can arrived at 91.33%.