Edge detection of potential field data with improved structure tensor methods

• A newly technique was presented to balance the edges.
• The new method can balance the amplitude well and get a higher resolution.
• A parameter p can avoid bring false edges and reduce the noise effect.

Edge detection is a requisite task in the interpretation of potential field data. There are many methods based on horizontal and vertical derivative of potential field data for edge detection and enhancement. The large eigenvalue of structure tensor can well delineate the edges of geological bodies, but it cannot outline the edges of small amplitude geological bodies clearly. In order to overcome this problem, this paper proposes three different normalization methods to improve the edge detection ability of the large eigenvalue of structure tensor, so that they can display the large and small amplitude edges simultaneously. Also, they do not produce additional false edges when real geological bodies contain positive and negative anomalies simultaneously. These methods were tested on synthetic and measured gravity gradient data and magnetic data. All of the results have shown that the new improved methods are effective for edge detection.

Edge results of traditional edge detectors THDR, Theta map, original structure tensor method and our new presented improved structure tensor methods. We can get that our new methods can balance the edges completely. The identified edges are clear and precise compared with original structure tensor method.Figure optionsDownload as PowerPoint slide