Statistical approach for microleveling of aerogeophysical data

In this paper, we present a new microleveling formula that uses statistical indicators to remove residual leveling errors commonly occurring in airborne data surveying. These line-to-line leveling errors which are not removed during regular data processing appear as short wavelength artifacts stretched along flight lines. Our approach aims to eliminate this residual noise from the grid of processed data by means of a moving square window. For each position of this window, we calculate the arithmetic averages and coefficients of variation in directions parallel and perpendicular to the survey lines, and estimate the noise level value. Moving the scanning window of one sample allows isolating a grid of noise and the correction is made by subtraction from the original data. To evaluate the efficiency of our method, we have tested it on a corrupted magnetic data with synthetic noise. We compared this technique with other methods and we found that the statistical microleveling is simple and produces very satisfying results.