摘要
提出一种基于傅立叶变换和模式识别技术的奥氏体超声探伤信号的缺陷定性分类方法。该方法利用快速傅立叶变换提取反映缺陷性质的特征值,然后运用BP神经网络对特征值进行缺陷定性识别。实验结果表明,此方法简单易行,且能较好地实现奥氏体的缺陷识别。
A signal recognition method for ultrasonic flaw detection of austenitic stainless steel is introduced based on Fourier Transform and pattern recognition. Characteristics are first extracted via Fast Fourier Transform （FFT）, and then Back-Propagation Network is used to identify the defect signal. Experiments show that the method is not only easy to implement, but also effective for identifying austenitic steel defect.
出处
《电子测量技术》
2006年第2期23-24,49,共3页
Electronic Measurement Technology
关键词
超声检测
神经网络
特征提取
ultrasonic testing, neural network, characteristics extraction.
