Feature Detection Using Sunshading

Abstract

Horizontal derivatives of potential field data are routinely used to sharpen the edges of linear features such as dykes and faults. They are also used as input to the sunshading filter. Given the azimuth and elevation of a source of illumination, this filter calculates the reflectance from a surface composed of the data to be interpreted. Linear features orthogonal to the illumination azimuth are enhanced, while those lieing parallel to it become less apparent. Since the sunshading filter is relatively insensitive to features of different azimuth from that being searched for, this paper first describes how to tune the angular sensitivity to any desired degree. Secondly, the field gradients are calculated in a cylindrical polar coordinate system rather than the usual Cartesian one. This allows both the enhancement of circular anomalies of any radii, and of linear features of any orientation. The cylindrical gradients may also be used with the sunshading reflectance algorithm. The new filters are demonstrated on various gravity and magnetic datasets.