Previously, this was done manually: it was difficult, time-consuming and expensive
Russian operator MTS announced that it was the first in the country to develop and launch a geoplatform for creating and updating digital 3D maps of the area using artificial intelligence. This solution will allow you to accurately and quickly select locations for installing base stations, which will improve the quality of mobile communications services and increase the efficiency of investments in infrastructure development, the company believes.
Previously, MTS purchased digital maps from external suppliers, and company specialists visually assessed topographical changes in the area and buildings based on satellite data and manually made changes to the digital model. Specialists must take into account all factors affecting signal propagation and the number of subscribers: the emergence of new residential areas and high-rise buildings, large shopping and sports facilities, industrial zones, the development of transport infrastructure and changes in the natural landscape. Do it manually — a long, labor-intensive and costly process.
With the launch of the new geoplatform, current high-resolution satellite images are processed by a neural network, which in seconds performs a large amount of work on visual analysis and three-dimensional identification of objects. As a result, vectorization of depicted objects is accelerated almost tenfold, MTS said. In this case, neural network algorithms automatically estimate the height of buildings and their configuration based on shadows or visible parts of walls to build a 3D model. As a result, planning a new MTS network or upgrading an existing one will proceed faster and more efficiently throughout Russia, the company notes.