摘要
针对传统小波变换中存在的边界问题,采用基于提升格式的小波变换,在数据内部及边界设计了相应的预测与更新算子,通过一个滑动的时间窗,使过程数据的去噪能够在线进行,同时加入粗大误差的去除,增强了算法的鲁棒性。通过数值仿真实验以及对实际过程数据去噪,表明该方法处理边界问题的有效性,能够达到较好的去噪效果。
This paper used the wavelet transforms based on lifting scheme for the boundary problem in the classical wavelet transforms. Pesigned the predict algorithm and update algorithm near and far from the boundary, which made it possible for online denoising of process data with a sliding window. And used the denoising of gross error to enhance the robust. The numerical experiment and denoising of process data show that the method can work well near the boundary of data and achieve better denoising result.
出处
《计算机应用研究》
CSCD
北大核心
2008年第10期3198-3200,共3页
Application Research of Computers
基金
黑龙江省科技攻关项目(GB06A106)
关键词
边界问题
提升格式
在线去噪
鲁棒性
boundary problem
lifting scheme
on-line denosing
robust