基于灰度-梯度共生矩阵和模糊核聚类的振动图形识别方法

PDF(1159 KB)

振动与冲击 ›› 2012, Vol. 31 ›› Issue (21) : 73-76.

论文

基于灰度-梯度共生矩阵和模糊核聚类的振动图形识别方法

丛蕊1 ,高光甫1 , 樊瑞筱1 , 乔磊1 , 张威1

作者信息 +

The method of vibrated image recognition based on gray-gradient co-occurrence matrix and kernel-based fuzzy clustering

Cong Rui1 Gao Guangfu1 Fan Rui xiao1 Qiao Lei1 Zhang Wei1

Author information +

文章历史 +

摘要

以往复机械振动参数图形为对象，提出了基于灰度-梯度共生矩阵和模糊核聚类的振动图形识别方法。利用灰度-梯度共生矩阵直接提取振动参数图形中的特征信息，将得到的纹理特征参量作为样本输入空间，通过Mercer核把输入样本映射到高斯特征空间后，在高维特征空间中进行聚类，从而实现往复机械故障智能诊断。实验结果表明，该方法可以获得较高的诊断精度，具有一定的可行性和有效性。

Abstract

This paper use the vibration parameters graphics of reciprocating machinery as the research object, this method was used to directly extract the vibration characteristics of the graphic information by gray - gradient co-occurrence matrix. Using gray-gradient co-occurrence matrix directly to extract vibration parameters of the graphic feature information, the texture characteristic will be inputted as the samples to space. Through the Mercer nuclear method, the characteristic parameters of the texture mapped to the Gaussian space, in order to have better clustering in the high-dimensional feature space, achieve intelligent diagnosis of reciprocating machinery fault. The experiment shows that this method has higher diagnosis accuracy, also has its feasibility and availability.

导出引用

丛蕊;高光甫;樊瑞筱;乔磊;张威 . 基于灰度-梯度共生矩阵和模糊核聚类的振动图形识别方法[J]. 振动与冲击, 2012, 31(21): 73-76

Cong Rui Gao Guangfu Fan Rui xiao Qiao Lei Zhang Wei. The method of vibrated image recognition based on gray-gradient co-occurrence matrix and kernel-based fuzzy clustering[J]. Journal of Vibration and Shock, 2012, 31(21): 73-76