摘要
文章将小波分析与竞争型神经网络相结合,构造了适应于零件分类的竞争型小波神经网络(CWN)。首先对零件先进行小波多尺度边缘检测,然后采用不变矩方法提取零件的特征,最后送入CWN加以分类。分类结果表明,CWN对相似零件能够有效分类,并且有一定的鲁棒性。
Based on the combination of wavelet resolution and competitive network, this paper proposes a competitive wavelet network which used for part clusteing. First the part edge is detected by multiscale wavelet,then the invariant moments is used to get the feature of the part,at last the CWN is used for clustering. The clustering result presents that CWN is good for clustering ,and the ability is robust.
出处
《组合机床与自动化加工技术》
2007年第3期94-97,共4页
Modular Machine Tool & Automatic Manufacturing Technique
基金
国家自然科学基金重点项目(60234010)资助
航空科学基金(05E52031)
南航创新实验室建设项目资助