摘要
为了减小轨检仪采集的轨道检测原始数据中夹杂的粗大误差噪声对检测结果的影响，提出经验模态分解法对检测数据滤波的可行性。对经验模态分解得到的第一层IMF1信号采用3σ准则识别粗大误差点并剔除，重构得到去除噪声后的信号。对限幅滤波法和经验模态法滤除噪声的波形进行分析和评价，进一步验证了经验模态分解法在处理非线性非平稳信号方面的优势。实例表明，论文提出的方法可以有效识别信号中的粗大误差点并剔除噪声信号，得到较为理想的滤波结果。
In order to reduce rail track detection of original data from the detector with bulky error noise influence on test results, empirical mode decomposition method is put forward in this article for the feasibility of detecting data filtering. The empirical mode decomposition gets the first layer of the signal which is IMF1 with 3σ criterion gross error identification point and eliminate, reconstruction after removing noise signal. The amplitude filtering method and empirical mode method are used to analyze and evaluate the waveform of noise filtering, to further validate the advantages of empirical mode decomposition method in nonlinear nonstationary signal processes . The examples show that the method put forward by the paper can effectively identify gross error point in the signal and eliminate the noise signal, the ideal filtering results are obtained.
出处
《自动化技术与应用》
2018年第2期97-101,共5页
Techniques of Automation and Applications
基金
国家自然科学基金（编号61572529）
关键词
去噪
经验模态分解
粗大误差
3σ准则
de-noising
empirical mode decomposition
gross error
3σ criterion
