摘要
为了简单高效地对电子元器件进行识别分类,本文提出了一种基于卷积神经网络的电子元器件的识别方法。该方法只需对图像进行简单的预处理,其网络模型能够自动提取图像特征,而且卷积神经网络能减少参数数量,降低计算的复杂度。实验结果表明,该方法无需对图像进行复杂的预处理,其网络模型能自动进行特征提取,能识别多种元器件,精度较高且复杂度较小,能够克服传统图像分类算法的诸多缺点。
In order to classify electronic components easily and efficiently,in this paper,a method for identifying electronic components based on Convolutional Neural Networks is proposed.The method does not require complicated image preprocessing steps.The convolutional network can automatically extract the image features.Moreover,the convolutional neural network can decrease the number of parameters involved in the calculation and reduce the computational complexity.Experimental results show that the method does not need to preprocess images complex.The network model can automatically extract features,identify a variety of components,with high precision and less complex,and can overcome the shortcomings of the traditional image classification algorithms.
作者
陈翔
俞建定
陈晓爱
翟影
CHEN Xiang;YU Jian-ding;CHEN Xiao-ai;ZHAI Ying(Institute of Information Science and Engineering,Ningbo University,Ningbo 315211,China)
出处
《无线通信技术》
2018年第2期7-12,共6页
Wireless Communication Technology
关键词
电子元器件
卷积神经网络
图像预处理
特征提取
图像分类
electronic components
eonvolutional neural networks
image preprocessing
feature extraction
Image classification
