摘要
应用磨粒形状特征参数、颜色特征参数和表面纹理特征参数对磨粒形态进行量化表征,并以此为输入矢量,引入径向基函数神经网络对磨损微粒进行自动分类识别,建立了适用于磨粒识别的径向基函数神经网络模型,并给出了具体算法.应用实例表明,径向基函数神经网络的收敛速度和识别率优于传统的BP神经网络.
A radius basis function (RBF) network was introduced to realize the automatic classification and recognition of wear debris, based on the qualitative characterization of the morphological features of the wear debris making use of the characteristic parameters of wear debris shape, color, and surface texture. Thus a neural network model based on the RBF network was established to classify and recognize the wear debris using those characteristic parameters as the input vectors. The algorithm of the established model was presented in detail as well. It was found that the neural network based on RBF was superior to conventional BP neural network in identifying and recognizing various wear debris. Namely, it had faster convergence speed and better accuracy than the recognition method based on BP neural network.
出处
《摩擦学学报》
EI
CAS
CSCD
北大核心
2003年第4期340-343,共4页
Tribology
基金
国家自然科学基金资助项目(50175069).