Rosleskhoz plans to use neural networks to predict forest fires forecasting and extinguishing forest fires. The relevant data are contained in the Rosleskhoz digital transformation program published on the FSIS KI portal.
In total, digital transformation will cost the department 6.7 billion rubles. It is planned to spend 35 million rubles on a remote sensing system for 100 percent of risk zones using AI in 2022 and 30 million in 2023. The creation of an automated system for forecasting the development of forest fires based on artificial intelligence will cost 30 million and 25 million, respectively. Cost indicators in the document are given in thousands of rubles.
In addition to the systems that will be used to predict forest fires and remote sensing, Rosleskhoz plans to digitize a number of state services related to the work of the department: forest inventory database with digitization of materials, creation of an automated systems for monitoring the execution of orders (SIOD Rosleskhoz) and the Forest Management Center.
At the same time, experts interviewed by Kommersant are not sure that Russia now has high-quality data sets on forest stands that could be effectively used by neural networks.
“The problem is data quality. For example, photoanalysis requires year-round images of areas from one equipment over several years. If you do it quickly, it can turn out like in the saying “Garbage in – garbage out” “, – said Nikolay Knyazev, the architect of Softline Digital.
At the same time, neural networks are in could help to simplify routine monitoring processes, says Dmitry Parshin, director of the Artezio development center (part of the Lanit group).
The area of forest fires in 2021 amounted to 18.2 million hectares, which is the greatest value for the entire 21st century. At the same time, according to Rosleskhoz, 18.1 million hectares burned down in 2020.
Climate change experts from the European Union Copernicus believe that forest fires in Siberia have become one of the most important sources of carbon dioxide emissions into the atmosphere in recent years .