The Thesis is dedicated to data-driven automation solutions aimed at improving the maintenance processes of overhead engineering infrastructure. A solution for implementing infrastructure inspection processes is developed and investigated, utilizing RGB and LIDAR datasets collected via unmanned aerial vehicles (UAVs) and leveraging machine learning for assessing the technical condition of infrastructure elements through large-scale data processing. The research is structured into five chapters, each addressing a critical aspect of modernization and analysis of power systems: a theoretical analysis of the structure of the power system and changes in its development trends, an overview of digital transformation in the energy sector, an investigation into the inspection processes of Latvia's electricity distribution system infrastructure, a theoretical evaluation of alternative methods for identifying visual and geospatial defects, and an experimental study on RGB and LIDAR data-based inspection solutions. Together, these components provide a comprehensive perspective on integrating new, more efficient methods into energy system management and fostering innovation within Latvia's power grid. The outcomes of the Thesis were developed in collaboration with JSC "Sadales tīkls", ensuring an industrial context for the research. The findings have been published in five scientific papers.