ONE-DIMENSIONAL INVERSION OF AN AIRBORNE TRANSIENT ELECTROMAGNETIC BY DEEP LEARNING BASED ON UNSUPERVISED AND SUPERVISED COMBINATIONS

A sample set of stratigraphic structure model airborne transient electromagnetic responses is established, the sample label is attached by unsupervised learning clustering technology, and the multilayer perceptron deep learning network with supervised learning is used to complete multiclassification tasks. Then, the sample set is input into the network for training to establish the inversion from the input response data to the output formation model. Verification results show that the prediction results are consistent with the types of sample stratigraphic models, which proves that the inversion method designed in this paper is correct, and efficient inversion from the test data to the prediction model is realized.