Detection of subsurface Qanats by Artificial Neural Network via Microgravity data


1 Instructor, Physics Department, Islamic Azad University, Najaf Abad Branch, Isfahan, Iran

2 Associate Professor, Earth Physics Department, Institute of Geophysics, University of Tehran and Center of Excellence in Survey Engineering and Disaster Management, Tehran, Iran

3 Professor of control, Electrical Engineering Department, University of Tehran, Iran

4 Associate Professor, Electrical Engineering, Electrical Engineering Department, Isfahan Technical University, Isfahan, Iran


A full automatic algorithm is designed to detect subsurface Qanats (sub terrains) via Artificial Neural Networks .We first gained the residual gravity anomaly from microgravity data and then applied it to a Multi Layer Perceptron (MLP) which was trained for the models of sphere and cylinder.
As a field example, the depth of a subsurface Qanat buried under the north entrance of the Geophysics Institute is determined through MLP (trained with noisy data).