As Elena Natarova, Marketing Director of Terra Tech JSC, stated while presenting this digital map in the Geospatial Artificial Intelligence section at the Innopolis AI Conference for Business (AI IN 2023), work on this project was fully completed by July 1, 2023.
The target audience of the project, as Elena Natarova reported, are federal and regional authorities, more than 1,200 users in total. According to her, already at the current stage of development, the digital map of economic activity allows real-time monitoring of forest resources, water surfaces, subsoil use and construction, which can be used by regional authorities for a comprehensive analysis of territories.
(AI) was actively used in the iceland whatsapp resource work on the project. She estimated the gain from using AI in processing Earth remote sensing (ERS) data at 97% in terms of time costs and 48% in terms of financial costs, with an error rate of no more than 3.5%. As a result, as Elena Natarova stated, the cost of data processing was 6 kopecks per 1 ha. AI was used to improve the quality of space images, for example, to eliminate extraneous objects - clouds, haze, etc.
"The presented digital map of natural resources and objects on a scale of all of Russia would have taken several years to prepare manually. It was neural networks that allowed us to make a cross-section of forested areas, logging, fires, construction, quarries and water throughout the country, which is more than 17 million square kilometers, in just one year. Artificial intelligence is one of the main drivers of the geotechnology market in the world. The Digital Earth - Services project demonstrated that we are at the forefront of automating the processes of analyzing space images and using advanced methods in implementing monitoring projects on a scale from a specific object to the entire country," Maxim Boltachev, CEO of Terra Tech JSC, assessed the importance of using AI in working on this project.
According to Elena Natarova, Terra Tech does not plan to stop there: the service's scope of application will be expanded, and new algorithms will be tested. The company also counts on new data from both Russian and foreign satellites.
Terra Tech also offers a commercial service, Pixel AI, which is free for plots up to 3,500 hectares. For now, as Elena Natarova noted, Pixel AI can be used to determine the boundaries of plots, but this service will be improved, and the nearest plans include creating a predictive analytics service for factors that can lead to a decrease in yield.
In general, according to other participants in the discussion, the use of artificial intelligence (AI) and machine learning allows to reduce the processing time of geodata up to 30 times, which greatly accelerates the output of products, given the large volume and labor intensity of geodata processing. Particularly acute, as emphasized by Mikhail Krinitsky, a researcher at the Shirshov Institute of Oceanology of the Russian Academy of Sciences, the task is during the processing of scientific research data: traditional processing of the results of one expedition takes an average of six months. The use of machine learning methods, according to him, made it possible to reduce this period to a week. At the same time, as Mikhail Krinitsky emphasized, the solution to this problem was quite difficult, and it took years: "Machine vision algorithms are imperfect and can hardly distinguish a log from a dolphin."
Konstantin Melnikov, Executive Director of Geocenter Consulting LLC, estimated the level of labor costs reduction in the development of geographic information systems (GIS) by three to four times. In his opinion, this is especially noticeable when updating the graph of the street and road network, which is the most complex and labor-intensive element in the creation of GIS products.
Elena Natarova particularly noted the fact that artificial intelligence
-
- Posts: 575
- Joined: Thu Jan 02, 2025 7:18 am