Research on the Universality of Convolutional Networks in Resistivity Inversion. Issue 1 (February 2021)

• Record Type: Journal Article
• Authors: Liu, Benchao; Guo, Qian; Pang, Yonghao; Jiang, Peng
• Is Part Of: IOP conference series. Volume 660:Issue 1(2021)
• Publication Date: 2021-02
• Subjects: Earth sciences -- Periodicals; Environmental sciences -- Congresses; Environmental sciences -- Periodicals; 550.5
• DOI: 10.1088/1755-1315/660/1/012060
• Languages: English
• ISSNs: 1755-1307

Abstract

Abstract: Resistivity inversion, as an important method to study the relationship between geological models and apparent resistivity data, is a typical non-linear problem. Convolutional neural networks have huge advantages in processing complex mapping relationships between images, so they are used to solve resistivity inversion problems. The convolutional neural network's weight sharing greatly improves the learning efficiency of the network, but there is a certain degree of incompatibility between this characteristic and the resistivity data model. In this work, the universality of the method was further verified by designing multiple complex anomalies and different background resistivities. The effectiveness of our proposed method is verified by comparing the inversion effects of different test sets with the results of traditional linear inversion.