基于改进半监督FCM聚类的复合材料超声检测脱粘缺陷识别
|
摘要
针对径向基(RBF)神经网络在进行超声检测脱粘缺陷识别时存在参数选择不确定、网络结构鲁棒性差等问题,提出一种改进的自适应半监督模糊C均值聚类(FCM)的RBF神经网络的方法,将kN近邻估计法和半监督模糊C均值聚类方法相结合,改进了隶属度函数,自适应确定聚类数目。将改进的RBF神经网络应用于超声检测脱粘缺陷识别,实验结果表明:与传统RBF神经网络相比,本方法减弱了孤立样本对网络结构的影响,增强了网络结构的鲁棒性,提高了脱粘缺陷识别的准确率,是一种较好的超声检测脱粘缺陷识别分类方法.
An improved radial basis function(RBF)neural network based on adapting semi-supervised fuzzy c-means is proposed to choose the parameters and has a robust structure.The method firstly combines the kN amphictyonic estimate with semi-supervised fuzzy c-means to improve the membership degree function;then determines the threshold to make sure the clustering members. The proposed approach is applied to recognize ultrasonic signals of un-bond defects,and the experimental results show that the proposed approach can lower the effect of the isolated point on the network structure and can be more robust while improving the accuracy rate.The proposed RBFNN approach is a feasible recognition algorithm for ultrasonic signals.
An improved radial basis function(RBF)neural network based on adapting semi-supervised fuzzy c-means is proposed to choose the parameters and has a robust structure.The method firstly combines the kN amphictyonic estimate with semi-supervised fuzzy c-means to improve the membership degree function;then determines the threshold to make sure the clustering members. The proposed approach is applied to recognize ultrasonic signals of un-bond defects,and the experimental results show that the proposed approach can lower the effect of the isolated point on the network structure and can be more robust while improving the accuracy rate.The proposed RBFNN approach is a feasible recognition algorithm for ultrasonic signals.
引文
[1]李惠芳,常宁.基于神经网络的金属与非金属材料粘接质量定量检测[J].北京工业大学学报,2009,35(8):1122-1125.
[2]DU Xiuli,WANG Yan,SHEN Yi.Ultrasonic signal classification based on ambiguity plane feature[J].Journal of Systems Engineering and Electronics,2009,20(2):427-433.
[3]王伯雄,刘文峰,柳建楠,等.基于特征的超声信号分类检测方法[J].清华大学学报:自然科学版,2012,52(7):941-945.
[4]徐超,王长龙,孙世宇,等.双小波神经网络迭代的漏磁缺陷轮廓重构技术[J].兵工学报,2012,33(6):730-735.
[5]张新贵,武小悦.基于RBFNN自适应混合学习算法的航天测控系统任务可靠性分配[J].航空动力学报,2012,27(8):1758-1764.
[6]师小红,徐章遂,敦怡,等.构件裂纹缺陷的超声识别[J].固体火箭技术,2007,30(6):556-558.
[7]徐桂云,蒋恒深,李辉,等.基于PCA-RBF神经网络的WSN数据融合轴承故障诊断[J].中国矿业大学学报,2012,41(6):964-970.
[8]周玉,朱安福,周林,等.一种神经网络分类器样本数据选择方法[J].华中科技大学学报:自然科学版,2012,40(6):39-43.
[9]李丹,顾宏,张立勇.基于属性加权的不完全数模糊C均值聚类算法[J].大连理工大学学报,2012,52(5):749-754.
[10]李昆仑,曹铮,曹丽苹,等.半监督聚类的若干新进展[J].模式识别与人工智能,2009,22(5):735-742.
[11]徐超,张培林,任国全,等.基于Parzen窗的油液原子光谱数据半监督FCM聚类研究[J].光谱学与光谱分析,2010,30(8):2175-2178.
[12]LI Dan,GU Hong,ZHANG Liyong,et al.A fuzzy cmeans clustering algorithm based on nearest neighbor intervals for incomplete data[J].Expert Systems with Applications,2010,37(10):6942-6947.
[2]DU Xiuli,WANG Yan,SHEN Yi.Ultrasonic signal classification based on ambiguity plane feature[J].Journal of Systems Engineering and Electronics,2009,20(2):427-433.
[3]王伯雄,刘文峰,柳建楠,等.基于特征的超声信号分类检测方法[J].清华大学学报:自然科学版,2012,52(7):941-945.
[4]徐超,王长龙,孙世宇,等.双小波神经网络迭代的漏磁缺陷轮廓重构技术[J].兵工学报,2012,33(6):730-735.
[5]张新贵,武小悦.基于RBFNN自适应混合学习算法的航天测控系统任务可靠性分配[J].航空动力学报,2012,27(8):1758-1764.
[6]师小红,徐章遂,敦怡,等.构件裂纹缺陷的超声识别[J].固体火箭技术,2007,30(6):556-558.
[7]徐桂云,蒋恒深,李辉,等.基于PCA-RBF神经网络的WSN数据融合轴承故障诊断[J].中国矿业大学学报,2012,41(6):964-970.
[8]周玉,朱安福,周林,等.一种神经网络分类器样本数据选择方法[J].华中科技大学学报:自然科学版,2012,40(6):39-43.
[9]李丹,顾宏,张立勇.基于属性加权的不完全数模糊C均值聚类算法[J].大连理工大学学报,2012,52(5):749-754.
[10]李昆仑,曹铮,曹丽苹,等.半监督聚类的若干新进展[J].模式识别与人工智能,2009,22(5):735-742.
[11]徐超,张培林,任国全,等.基于Parzen窗的油液原子光谱数据半监督FCM聚类研究[J].光谱学与光谱分析,2010,30(8):2175-2178.
[12]LI Dan,GU Hong,ZHANG Liyong,et al.A fuzzy cmeans clustering algorithm based on nearest neighbor intervals for incomplete data[J].Expert Systems with Applications,2010,37(10):6942-6947.