Depth Recognition of Underground Power Cables Based on Self-attention Mechanism and LSTM Network

In view of the problems of slow acquisition speed, few elements, and low accuracy of traditional ground detection methods currently used in our country. At the same time, in view of the long and short-term memory network (LSTM) in deep learning, it has the advantages of automatic feature extraction and integration of classification and recognition. Based on the research foundation of transient electromagnetic method (TEM), this paper proposes a deep recognition method of underground power cables based on self-attention mechanism and LSTM network. First, the induced voltage at different time points is normalized, and the TEM apparent resistivity is quickly obtained through the LSTM-Self-Attention network. Among them, the LSTM-Self-Attention network weights are optimized by self-attention mechanism to improve the accuracy of power cable depth recognition. Finally, the power cable depth identification method proposed in this paper is simulated, and the experimental results verify the effectiveness of the proposed method.