中文摘要：

基于动力系统模型的局部投影非线性降噪算法，对混沌信号和一大类非平稳信号具有明显的降噪效果。利用局部投影算法对信号进行降噪处理，邻域半径的选择对降噪结果有很大的影响。我们研究了基于递归分析的邻域半径参数的估计和基于噪声强度的邻域半径参数估计。通过对Logistic映射和Lorenz模型叠加不同大小的高斯白噪声，对两种邻域半径估计方法进行分析比较，得到了基于噪声强度的邻域半径参数估计方法具有更好的降噪效果，为局部投影降噪算法的应用奠定了基础。

英文摘要：

The algorithm of local projective nonlinear noise reduction based on nonlinear dynamic system is more effective for noise reduction of chaotic signal and non-stationary signals. In local projective nonlinear noise reduction algorithm, the selection of neighborhood size parameter has an important effect on the result of noise reduction. Two methods of how to choose neighborhood size parameters were discussed. The methods are based on recurrence plot and noise level estimation respectively. After adding write noise of different level to Logistic map and Lorenz model, the noise reduction results were obtained which are more satisfactory compared with the results from other noise reduction methods, And the results also show that the neighborhood size determined by noise level is superior to that determined by recurrence plot.