Removing coherent disturbance by mean value weighting

This paper presents a data processing method for nois e removal which ca n decrease the damage to the effective waves to the minimum. For signals in freq uency segments without coherent disturbance the wavelet frequency-division is a d opted to remove noises. For signals with coherent disturbance the window-divisi o n is adopted to make the noise removal processing performed only within segments with disturbance. The picking of coherent disturbance is based on the mean valu e weighting technique, which is new, precise, and practical for noise picking. R eal data processing demonstrates that the method can eliminate coherent disturba nce while fully preserving effective waves from loss and its negative effect is very small.