超声信号的特征提取与选择在缺陷分类识别中的应用

摘要

It is very important to extract and select the feature of ultrasonic flaw echo in ultrasonic detection. This paper makes a study of the feature extraction of ultrasonic flaw echo in time and frequency domain according to the signal characteristics,presents the combination algorithm based on SBS and SFS and analyzes the proposed method. It shows that the dimension is reduced effectively and good efficiency is got.%超声缺陷回波信号的特征提取与选择是超声波检测的基础和关键.结合待检测超声信号的特点,在时域、频域上研究了超声缺陷信号的特征提取,提出SBS和SFS结合的特征选择方法.在满足识别准确率的同时,有效地降低了分类器输入向量的维数,提高了运算效率.