| 基于BP神经网络的工业视觉检查系统研究 |
| |
| 引用本文： | 杨玲,刘桂艳,姜立芳,于晓洋. 基于BP神经网络的工业视觉检查系统研究[J]. 哈尔滨工业大学学报, 2004, 36(6): 845-847 |
| |
| 作者姓名： | 杨玲 刘桂艳 姜立芳 于晓洋 |
| |
| 作者单位： | 哈尔滨理工大学,测控技术及通讯工程学院,黑龙江,哈尔滨,150040;东北轻合金有限责任公司铝品制造厂,黑龙江,哈尔滨,150040 |
| |
| 摘 要： | 研制了一套能在生产线上自动检查产品形状、尺寸及缺陷的工业视觉检查系统．在模式识别阶段采用BP神经网络，并针对BP算法的局限性，给出了一种改进的BP算法．采用有导师学习方式，并选取一种新的误差函数，根据训练过程中的样本情况，动态、系统地调整其参数．实验结果表明，该系统可正确识别工业生产线上的零件，有效地提高生产率和保证产品的质量．
|
| 关 键 词： | 图像处理 BP算法 神经网络 工业视觉 模式识别 |
| 文章编号： | 0367-6234(2004)06-0845-03 |
| 修稿时间： | 2003-11-20 |
Industrial vision examination system based on BP neural network |
| |
| YANG Ling,LIU Gui-yan,JIANG Li-fang,YU Xiao-yang. Industrial vision examination system based on BP neural network[J]. Journal of Harbin Institute of Technology, 2004, 36(6): 845-847 |
| |
| Authors: | YANG Ling LIU Gui-yan JIANG Li-fang YU Xiao-yang |
| |
| Affiliation: | YANG Ling~1,LIU Gui-yan~2,JIANG Li-fang~1,YU Xiao-yang~1 |
| |
| Abstract: | An industrial vision examination system that can automatically examine space, size and disfigurement of products in industrial production line is introduced. BP neural network is adopted to accomplish the computer image identification system. An improved BP algorithm is applied in the identification of industrial tarts, which has a tutor style to study. According to the status of samples in the course of training, a new error function to adjust its parameter dynamically and systemically is adopted. Experimental results indicate that this system can improve the productivity and guarantee the quality of products. |
| |
| Keywords: | image processing BP algorithm neural network industrial vision pattern recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
