Image enhancement of aeromagnetic data using automatic gain control

Abstract

Image enhancement consists of raising subtle trends above the detectability threshold usually either by increasing the relative amplitude or by emphasising the continuity of the feature. Images of the vertical magnetic gradient are in themselves an enhancement of the shallow source anomalies present in total magnetic intensity data. Even in the gradient image, low amplitude anomalies may be obscured by intense anomalies which may be several orders higher in amplitude. One way to overcome this is to sacrifice amplitude fidelity, retain anomaly shape, and enhance the gradient data using automatic gain control (AGC). By definition automatic gain control implies that for a given frequency, output amplitude is independent of input amplitude. The controlling parameters when applying automatic gain control are the window size over which the gain is computed and the gain function which controls the relative amplification of signal and noise. Since gradient data are only rarely collected during an aeromagnetic survey, the creation of an AGC-enhanced gradient image can be achieved in different ways. The computation of the vertical gradient and subsequently of the AGC can be carried out either on profile data or on gridded data. The results are similar but the noise level and the amount of high-frequency information in the output may be significantly different depending on the order of the processing. Experiments using high-resolution aeromagnetic data indicated that for the purposes of trend identification, the image derived from gridded AGC-enhanced profile data was preferable to that obtained from AGC-enhanced grid data. A suitable gain function was the inverse root mean square value computed within a sliding window. The image could be significantly improved by setting the base-level of the gradient profile data to zero before applying the AGC.