摘要 : Extracting the edges of gravity and magnetic data has been commonly used as an important tool for determining the lineaments of buried structures. There are many edge detectors based on potential field derivatives for highlighting... 展开 Extracting the edges of gravity and magnetic data has been commonly used as an important tool for determining the lineaments of buried structures. There are many edge detectors based on potential field derivatives for highlighting the source edges. The universal restrictions of these detectors are that they cannot produce balanced images, or the results obtained from use of these filters have low resolution. To solve these problems, I introduce a new filter that uses the Heaviside step function of the vertical derivative of the potential field data to detect the source edges. The novel filter is demonstrated on synthetic and real gravity datasets over the Witwatersrand basin (South Africa) and the estimated edges are compared with those determined from popular detectors. The findings show that the proposed method had better performance than other detectors in outlining the geologic features which were not seen in the observed fields. 收起