【作者】 周小波； 张贤达；

【Author】 Zhou Xiaobo Zhang Xianda State Key Laboratory of Intelligent Technology and Systems Department of Automation,TsingHua University,Beijing,100084,China.

【机构】 智能技术与系统国家重点实验室清华大学自动化系；

【摘要】 在信号去噪声中,要同时去除脉冲噪声和白噪声,并保持信号的突变现象,有一定的困难。本文研究了非参数分析在去除白噪声的同时也能去除脉冲噪声。我们利用非参数分析的多项式回归和核估计量对信号做平滑和预测处理,再联合交叉有效平滑思想提出了基于非参数分析的一类非线性滤波。并通过摸拟试验检验了该方法,并与有限混合中位数滤波(FMH)法和自适应小波收缩算(WSA)法做比较,试验结果说明了该方法是切实可行的,并优于FMH和WSA。更多还原

【Abstract】 In noise attenuation,it is difficult to reduce impulse noise and white noise and retain the jump phenomen at same time.In this paper,we study the problem by nonparametric analysis.Firstly, we use smoothing and predicative filter to process the signal via multipolyial regression and kernal estimator, then we give the nonlinear filters based on nanparametric by combining Cross-Validation.At last,we do some examples to test the phcv method,and make a comparison with FIR-Median Hybrid(FMH) and Adaptive wavelet shrinkage algorithm(WSA).The ersult of examination is shown our method is a effective method and better than the FMH and WSA.更多还原