Detection and Removal of Noisy Land Magnetic Data Spikes

Detection of land magnetic data spikes is an important issue in magnetic data processing. In addition, the presence of noisy contributions such as spikes, stripes and zigzag effects in magnetic data visualization represents the most common flaw that may degrade the image. Rendering the correct detection and identification of features very uncertain. In this study, a script called "Window_Despike" was written to mark the spike data points and give the index of each spike point. Spikes were replaced by linearly interpolating the adjacent "good" amplitudes, or replaced by Not a Number (NaN). Different windows size starting from window size of 3 till 9 compared and found that the best window size is 5.