Edge detection and recognition based on structure tensor
Wen‐na Zhou · 2011
Edge detection and enhancement techniques are popularly used to outline edges of causative sources with potential field data. In this paper, a new edge finder algorithm is proposed initially based on structure tensor. The structure tensor is established originally from the field gradient tensor, but is accurate, resistant to random noise, and has high resolution features. There are different types of gravity synthetic models created for testing the validity of this method. The test shows that the method is more robust than the conventional methods.