| 涡流检测信号输入神经网络前特征值提取问题的论证 |
| |
| 引用本文： | 钟洁 王素菊. 涡流检测信号输入神经网络前特征值提取问题的论证[J]. 无损检测, 1999, 21(3): 97-98,114 |
| |
| 作者姓名： | 钟洁 王素菊 |
| |
| 作者单位： | [1]南京建筑工程学院 [2]南京航空航天大学 |
| |
| 摘 要： | 将神经网络技术应用于涡流检测信号的自动识别与分类是目前无损检测界关注的问题。针对涡流检测信号输入神经网络前特征值得提取问题研究了傅里叶描述符法的应用可行性并作了大量实验验证。
|
| 关 键 词： | 信号处理 涡流检测 神经网络 特征值提取 钢管 |
PROOF OF THE EXTRACTION OF EDDY CURRENT SIGNAL CHARACTERISTICS BEFORE THEIR INPUTTING INTO NEURAL NETWORK |
| |
| Zhong Jie Wang Suju. PROOF OF THE EXTRACTION OF EDDY CURRENT SIGNAL CHARACTERISTICS BEFORE THEIR INPUTTING INTO NEURAL NETWORK[J]. Nondestructive Testing, 1999, 21(3): 97-98,114 |
| |
| Authors: | Zhong Jie Wang Suju |
| |
| Abstract: | Close attention is paid on the application of neural network technology to automatic recognition and grading of eddy current signals in the field of NDT nowadays. The feasibility of the application of Fourier descrip-tors to the extraction of eddy current signal characteristics before their inputting into neural network has been studied and proved by experiments. |
| |
| Keywords: | Eddy current testing Signal processing Characteristic |
| 本文献已被 维普 等数据库收录！ |
