Digital transformation of regional agricultural GIS carried out in the Republic of Tatarstan

2023: Digital Transformation of Regional Agricultural GIS

InnoGeoTech On July 13, 2023, the company "" shared TAdviser information with one of its largest and most complex projects in terms of project development - the case of creating a geo-information module of the agro-industrial sector Republic of Tatarstan for the needs of the regional. According Ministry of Agriculture to representatives of InnoGeoTekh, work on the project was carried out during 2022, and the launch of the GIS module took place in 2023.

Spatial data on the Republic of Tatarstan is now available in one window mode

As the name suggests, the module is responsible for working with geodata in crop production - these are millions of hectares of arable land, tens of thousands of fields, almost a thousand agricultural producers (from small agricultural farms to the largest agricultural holdings) and a huge array of geodata. For this reason, the key demand for a digital transformation of an industry solution of this nature is the transition from manual field data to automated, as well as the introduction of a single window for working with this data for a number of employees - from line personnel to manager.

By the start of the project, the InnoGeoTech team had already accumulated a comprehensive expertise in the digitization and vectorization of large geospatial data and the use of ML technologies to work with them. The company developed digital counterparts of cities, monitored forest and agricultural lands, land use facilities.

Further, based on the needs of the end users of the system and their own competencies, a backlog of 3 key problems was formed, the solution of which will be required within the GIS module. Certification of all fields in the region, that is, collection and processing directly from agricultural producers (SCTP) of the following data: field contour, growing crop, crop rotation, etc. Visualization of the obtained data in one window mode for prompt access to information of employees of the republican Ministry of Agriculture and related departments.

Verification (verification) of field data using Artificial Intelligence (machine vision) technologies to analyze space images and determine the fact of plowing.

The project team immediately realized that it was necessary to implement a GIS module of such a configuration that had not yet been implemented in any region of the Russian Federation. It was necessary not only to create a tool for collecting data on the fields of the region, but also to develop a number of features to optimize this work, share team work and generate automatic reports on various user requests.

In the development of this project, the team's experience in creating its own web-gis "GeoHub" came in handy. Competencies in the implementation of a powerful geo-information web-gis platform helped to form an equally functional regional industry system. In parallel, the development team created more than 40 utilities and tools for automated work with the import and export of geo-information data, automatic topological checks, integration with many information systems and modules, the formation of a set of online reports and the implementation of the role model.

Admin account with all fields in the region for quick access to up-to-date information

In my opinion, such solutions and systems should be in every region and in every industry. The volume of accumulated unenciphered data and the colossal work that is tied to these materials should be translated "into digital" as early as possible and as high-quality as possible, more accessible. Digitalization is an integral part of the modern world, and the steps that are being taken for July 2023 are not even an innovation and perspective, but the need and the main condition for normal work in various sectors of the modern state. The sooner business and the state realize the advantages of this approach (which should be product and capital, supported and modernized for several years, and not fast and project, "made and gave") and allocate funds for such development, the faster they will receive feedback, including in the form of finance and resource savings. told Alexei Mirolyubov, project manager of InnoGeoTekha

The primary task solved within the framework of the gis module is the certification of fields, that is, the collection and storage of data on the fields of the region directly from agricultural producers (SKTP) in one system. The user interface was designed so that each farmer had his own personal account, an online version of his fields with real size, the type of crops grown and other data. Farmers make all these changes on their own.

For further work, it was necessary to automate the collection of this information, in the case when the CFS already has a layer with fields in vector format - geojson, shp, etc. If for small agricultural enterprises it is not difficult to manually draw their fields, then for large agricultural enterprises and holdings with thousands of fields, entering information using manual geometry rendering was very labor-intensive.

The system also contains data from different periods: when you next enter information, you do not have to draw fields from scratch - it is enough to duplicate the last season to another, making the necessary changes, for example, in terms of crop rotation. As of July 2023, information on almost 3 million hectares of land from more than 900 agricultural enterprises of Tatarstan was registered in the Land Passport module. About 92% of all farmland in the region is stored digitally.

An important task was also the introduction of specialized neural networks for verifying the certification data received from the CSTP. Analyzing satellite images of average spatial resolution, neural network algorithms learned to recognize the fact of plowing fields, find wooded areas, and conduct spatial analysis to identify unregistered areas. The information received is used by employees of the Ministry of Agriculture of the region in operational activities and the distribution of subsidies to the agricultural industry.

Over the past 5 years, the team has been developing and optimizing models for segmentation, classification, detection of land use in various regions of the country (agriculture and forestry, subsoil use, urban studies). The company writes models of neural networks on its own. However, it uses time-tested AI architectures, such as Unet and DeepLab. The results of neural networks solve a large number of urgent problems related to updating, accumulating data, and most importantly faster than by the hands of cartographers.

Specialized neural networks carry out automated monitoring of arable land

The GIS module was tested in 2022 and is already in full operation by the state authorities of the Republic of Tatarstan. A project of a similar scale may affect the digital transformation of the agricultural industry of the regions of the Russian Federation in the context of working with data.