3-D Gravity Anomaly Inversion Based on Improved Guided Fuzzy C-Means Clustering Algorithm

Abstract

The geophysical inversion with combining prior information is very important for resource exploration and studies of the Earth’s internal structure. Guided fuzzy C-means clustering inversion (FCM) is normally applied for the Tikhonov regularized inversion, but has the shortcoming of uniform model parameter shrinkage, leading to inaccuracy. In this paper, an improved guided fuzzy clustering algorithm is proposed by adding a fuzzy entropy term to the original guided FCM. This method not only enforces the discrete values to a high degree of approximation by guiding the recovered model to cluster tightly around the known petrophysical property values, but also calculates the distributed characteristics of the model parameter set. Based on this method, the shortcoming of uniform shrinkage of the original guided FCM clustering algorithm is improved, and more accurate inversion results are obtained, making the FCM method more efficient and broadly applicable. Furthermore, a new parameter search algorithm is proposed to accelerate the search speed. The results recovered by using this method with three kinds of theoretical gravity anomaly data show more accurate density anomalies compared with the results generated from the original guided FCM clustering inversion and greater efficiency in the parametric search process when using the new parameter search algorithm. The improved FCM clustering algorithm could enable more extensive and efficient use of gravity inversion.