Detection of high voltage transmission towers using Sentinel-1 satellite images with a machine learning approach based on remote sensing methods

Electric energy plays a critical role in the economic and technological development of modern societies. By enabling uninterrupted service delivery in key sectors such as industrial production, healthcare, education, and communication, electricity enhances the quality of life. The transmission of electric energy is carried out through high-voltage lines to ensure efficient energy distribution. These lines minimize energy transmission losses, thereby increasing energy efficiency and sustainability. The safe and continuous transmission of electric energy is of great importance for the robustness and maintenance of infrastructure. In this study, it is aimed to detect high-voltage transmission line poles around Batman using remote sensing and different machine learning techniques. The Sentinel-1 Synthetic Aperture Radar (SAR) satellite was used. DY (Vertical-Horizontal), DD (Vertical-Vertical) polarization modes formed the dataset. These datasets were analyzed by algorithms belonging to the supervised learning model type, namely Support Vector Machine, KNN, Decision Tree, Quadratic Discriminant, Naive Bayes. As a result, the Support Vector Machine model achieved an accuracy rate of 85.0%; Quadratic Discriminant model 82.5%; KNN model 82.2%; Decision Tree model 76.8%; and Naive Bayes model 74.0%. The study allows for the more accurate, faster, and periodic control of high-voltage poles using artificial intelligence. Considering that traditional methods for inspecting high voltage lines are time-consuming and costly, the aim of this study is to contribute to the monitoring and improvement of maintenance processes of energy transmission lines, thereby enhancing the efficiency of energy infrastructure.