摘要
针对焊缝特征提取中普遍采用的线段拟合方法计算量大的问题 ,探讨了基于小波变换的焊缝特征提取方法 ;将离散小波变换和奇异分析理论引入焊缝特征提取问题 ;利用多尺度小波分析计算量小以及对信号奇异点敏感的特征 ,采用db2小波对焊缝特征进行分析 ,通过采用小波变换模极大值传播规则 ,大大减小了计算量 .实验证明此方法具有较好的性能 ,有效地排除了局部信号对焊缝整体特征的影响 。
Presents the development of a welding seam position information extraction algorithm based on wavelet analysis to solve the problem of high computation cost with the widely used linear regression method with the discrete wavelet transform and singularity detection theory introduced for welding seam postion information extraction,and the db2 wavelet adopted to analyze the characteristics of welding seam signal because multi resolution wavelet analysis has low computation cost and is sensitive to singularity of signal.and the propagation rules of wavelet transform extrema proposed to decrease the computation cost greatly and concludes from Experiments that the method proposed has good performance,local sgnal's affection on global characteristic of the signal is effectively excluded,and the welding seam information can be quickly identified.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2001年第3期389-392,共4页
Journal of Harbin Institute of Technology
关键词
焊缝识别
小波分析
奇异分析
焊缝位置
wavelet analysis
welding seam recognition
singularity detection
