中值约束下的矢量分解去噪

Noise elimination with median-constrained vector resolution.

首先分析了中值滤波去噪与矢量分解去噪的特点 ,然后给出了将两种方法结合的新的去噪方法———中值约束下的矢量分解去噪技术。该项技术先对要处理数据中的样点进行记录 ,沿信号方向求取中值 ,并以中值与该样点组成一矢量 ,然后利用矢量分解的办法在定义的标准矢量上投影分解 ,从而达到去噪的目的。理论记录与实际资料处理表明 。

In this paper the behaviors of median filtering and vector resolution for noise elimination are analyzed, and a new method, called noise elimination with median constrained vector resolution, is presented by combining the above methods. The new method first deals with sample records, then calculates median values along the signal direction and combines the resulted median values with the sample to form a new vector, and then finally projects the new vector to a predefined standard vector according to the principle of vector resolution. Thus the goal of noise elimination is realized. The results of synthetic and real data processing show that the proposed method can satisfactorily eliminate noises.