| 基于神经网络的注塑制品表面缺陷自动识别方法的研究 |
| |
| 引用本文： | 刘斌,孙加伟,吴盛金.基于神经网络的注塑制品表面缺陷自动识别方法的研究[J].塑料工业,2012,40(2):38-41. |
| |
| 作者姓名： | 刘斌 孙加伟 吴盛金 |
| |
| 作者单位： | 华南理工大学聚合物成型加工工程教育部重点实验室聚合物新型成型装备国家工程研究中心,广东广州,510640 |
| |
| 摘 要： | 将基于误差反向传播算法(BP)的神经网络引入到注塑制品表面缺陷的自动识别.介绍了如何选择合适的BP神经网络,包括网络层数的选取、学习算法的选取等.最后分别利用90组样本对BP神经网络进行训练和仿真,得到制品表面缺陷的平均识别率达84.44％,说明利用BP神经网络对于注塑制品表面缺陷进行识别是可行的.
|
| 关 键 词： | 注塑制品 表面缺陷 自动识别 神经网络 |
Automatic Recognition of Surface Defect on Injection Product Based on Neural Network |
| |
| LIU Bin
,
SUN Jia-wei
,
WU Sheng-jin.Automatic Recognition of Surface Defect on Injection Product Based on Neural Network[J].China Plastics Industry,2012,40(2):38-41. |
| |
| Authors: | LIU Bin SUN Jia-wei WU Sheng-jin |
| |
| Affiliation: | (The Key Laboratory of Polymer Processing Engineering of Ministry of Education,National Engineering Research Center of Novel Equipment for Polymer Processing,South China University of Technology,Guangzhou 510640,China) |
| |
| Abstract: | The error back-propagation algorithm(BP)neural network was applied into the automatic recognition of surface defects of injection product.Select ways on the appropriate BP neural network were introduced,including the selection of the network layers and learning algorithms,etc.By using ninty samples to the training and simulation of BP neural network respectively,the average recognition rate of surface defects could reach 84.44%.It demonstrated that the use of the BP neural network to recognize surface defects of injection product was feasible. |
| |
| Keywords: | Injection Product Surface Defect Automatic Recognition Neural Network |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
