paper with LaTex

xiao,jia xiao*
China University of Geosciences, Wuhan, 430074, China


In this paper, eigenvalues of the full gravity gradient tensor (GGT) are used to detect edges of geological structure. First, the solving of GGT eigenvalues is discussed; then a new edge detection method is proposed by using the eigenvalues of GGT. Comparing with the pervious edge detection method based on curvature gravity gradient tensor (CGGT), the full gravity gradient tensor containsmore independent gradient components that are helpful to detect more subtle structures of the sources. The proposed method is applied to the synthetic data with and without noise to determine the locations of the edges of the mixed positive/negative contract density bodies. It has also been tested on real field data. All of the experimental results have shown that the newly proposed method is effective for edge detection.


Edge detection is required for interpretation of gravity field data,and has been widely used in tectonic studies and mineral explorations(Paterson and Reeves, 1985). To gain a better understanding of gravity sources, a variety of methods based on the use of directional derivatives of different orders have been developed to locate anomalous source boundaries (Miller and Singh, 1994; Verduzcoet al., 2004; Cooper and Cowan, 2006; Wijns et al., 2005; Ma and Li,2012; Foks and Li, 2013).
In this paper, we propose an edge detection method based on the analysis of eigenvalues of GGT. Considering the actual geological situation, the positive and negative density bodies usually coexist simultaneously. The proposed method has designed to outline the positive/negative density bodies with the corresponding values.

Full gravity gradient tensor and its eigenvalue analysis

Theory of GGT


The practical gradio meter systems for rapidly measuring GGT have been developed (Bell et al., 1997; Jekeli, 1993). Mikus and Hinojosa (2001) propose the method of approximating of the GGT from measured gravity data. The full gravity gradient tensor, symbol T, can be written in the form : T = [\(\frac{\partial^{2}U}{\partial_{x}^{2}}\) \(\frac{\partial^{2}U}{\partial_{x}\partial_{y}}\) \(\frac{\partial^{2}U}{\partial_{x}\partial_{x}}\)\(\frac{\partial^{2}U}{\partial_{y}\partial_{x}}\) \(\frac{\partial^{2}U}{\partial_{y}^{2}}\) \(\frac{\partial^{2}U}{\partial_{y}\partial_{z}}\)] = [11-11121202]

\begin{equation} \label{eq.1} \label{eq.1}p=g_{xx}g_{yy}+g_{xx}g_{zz}+g_{yy}g_{zz}−g_{zx}g_{xz}−g_{xy}g_{yx}−g_{zy}g_{yz};\\ \end{equation}

kankanba haikeyi a a a a a



sdfa ] Set out the conclusion of this original[LABEL:eq.1] software publication.



Optionally thank people and institutes you need to acknowledge.