The paper is concerned with justification and demonstration of the new method for parameter determination for induction sensors used in overhead electric power line monitoring systems. The research focuses on using artificial intelligence techniques, namely, convolutional neural networks, to identify the signal information parameters in order to overcome problems connected with high sensitivity of induction sensors to electromagnetic noise. The paper also describes in detail the process of signal analysis and filtering using a convolutional network, which makes it possible to improve the measurement accuracy even under the conditions of high-intensity noise. The developed neural network architecture and its training method are also described; the training results prove high efficiency of the proposed method in actual operating conditions. The paper emphasizes significant contribution of artificial intelligence techniques into energy infrastructure reliability and security improvement especially in the field of early-warning failure detection. It is considered in detail how deep learning can improve the processes of induction sensors monitoring offering a practical solution for monitoring capability enhancement. This research work both shows the artificial intelligence potential in power engineering and provides the foundation for development of more reliable and efficient monitoring systems contributing to continuous power supply and it also emphasizes the importance of integrating new technologies into traditional electric power systems to improve their stability and provide higher security.