人工神经网络在金属餐具缺陷检测中的应用

Application of artificial neural network in the defect detection of metal tableware

针对传统金属餐具缺陷检测主要依靠于人工,导致检测成本高、效率低的问题,提出一种采用机器视觉与人工神经网络相结合的金属餐具缺陷检测方法。以金属勺子为例,对勺子缺陷进行分析、特征提取,建立图片特征库,构造了人工神经网络,并通过对神经网络的训练和检验,实现了对缺陷的自动检测。

Traditional metal tableware defects detection mainly relies on manual work,which makes the detection cost high and efficiency low. To solve this problem,the combination of machine vision and artificial neural network is adopted to realize automatic defect detection. Taking the metal spoon as an example,the image feature library and the artificial neural network are constructed by analyzing and extracting the features of metal spoon,and the automatic defect detection is realized through the training and inspection of the network.