Satellite Imagery Analytics In Environmental Monitoring

On April 23-28, 2023, EOSDA scientists Olena Kavats, Dmitriy Khramov, and Kateryna Sergieieva will attend the EGU General Assembly conference and present their achievements in the GoldenEye EU H2020 project. The team will discuss geoscience from the perspective of space monitoring with their report “Monitoring Active Mining Areas in Operation using Sentinel-1 Coherence Time Series”.

EOS Data Analytics has been participating in the three-year GoldenEye program since May 2020, together with 17 partner organizations from 5 countries in Europe. The project is aimed at creating a single information space for comprehensive monitoring of underground and open pit mining processes. A broader long-term impact includes assistance in the implementation of environmental monitoring for sustainable mining practices.

We reached out to Olena Kavats, Data Scientist and the GoldenEye Project Coordinator at EOS Ukraine, a subsidiary company of EOSDA, to check in on the GoldenEye project and ask about the nearest plans of the science team.

How Challenging Was It For Your Team To Land This Project Considering Its Complexity?

The framework agreement between Ukraine and the EU opens opportunities for the participation of Ukrainian enterprises in programs like Horizon 2020. The GoldenEye project was launched as a result of the European Commission awarding an H2020 grant opportunity to the GoldenEye consortium members, including us.

Taking into account our experience in the field of environmental protection, we could apply our expertise in using satellite imagery and data analytics to facilitate environmental monitoring.

What Has Been EOSDA’s Contribution To The GoldenEye Project So Far?

One of the most exciting challenges of the project was the need to improve the spatial resolution of the land surface temperature data. The original satellite images from Landsat-8 and Landsat-9 don’t give as high resolution as we needed for accurate temperature monitoring, especially for the small field trial sites that we were given. To achieve that, we improved an ElasticNet algorithm to downscale the data to a 10-meter spatial resolution.

This regression model became a great backbone for building highly detailed maps of surface temperature, thereby being able to assess the impact of a mining site on the environment. The advantage of the GoldenEye project is that we have a chance to prove the efficiency of our methods during the field trips and ground data collection at the actual mining sites.

We published a scholarly article ^{Kavats, O., Khramov, D., Sergieieva, K., & Vasyliev, V. (2020). Algorithm for statistical downscaling of the land surface temperature using ElasticNet. Review of the Bulgarian Geological Society, 81(3), 195-197.} if anyone’s interested in how we improved ElasticNet for the goals of geospatial analysis.

Beyond surface temperature and humidity monitoring, we’ve also been working on satellite data processing, as well as a number of other AI tools. For instance, we were able to facilitate the remote assessment of the slope stability, the vegetation of adjacent recultivated territories, and the change dynamics in water surface areas of the tailings dams ^{Kavats, O., Khramov, D., Sergieieva, K., Surface Water Mapping from SAR Images Using Optimal Threshold Selection Method and Reference Water Mask, Water 2022, 14(24), 4030.}.

The Project Plan Includes Validating The Technology In 6 Field Trials. At Which Stage Is It Right Now?

Right now, we’re in the middle of implementing the results of environmental monitoring into the platform. We already found a way of analyzing the satellite data that we needed for achieving the environmental monitoring goals. Now we are validating the obtained results in field trials. In the future, we are expecting the AI platform that we’re working on to operate on a continuous basis and be available for non-academic professionals. It is planned to open the platform for use by large mining and processing enterprises.

Is Participating In The GoldenEye Bringing EOS Data Analytics Closer To Sustainable Development Goals?

It definitely is! The Horizon 2020 program gave our data scientists and geologists the opportunity to develop and prove their methods in real-life conditions. By implementing the improved data processing and predictive algorithms, we can provide the necessary information that will help to solve pressing issues like evaluating the integrity of currently operating tailing dams as well as ones in post-closure phases, monitoring the disposal of production waste, managing environmental risks assessments, predicting landslides, and much more.

EOS Data Analytics is committed to long-term sustainable development goals outlined by the UN. In particular, we align with 10 out of 17 United Nations Sustainable Development Goals. Participation in the H2020 GoldenEye project allowed us to bring forth practical solutions to goals like Climate Action, Industry Innovation and Infrastructure, and Responsible Consumption. The GoldenEye platform will facilitate more efficient change management in mining sites while promoting resilient infrastructure as well as sustainable production and consumption patterns.

When it comes to the mining industry, there’s a variety of hazards to be aware of during risk management. Those could be spontaneous explosions, dam leaks, quarry wall collapses, and mining waste pollution, to name a few. But there is even more potential damage than the eye can see. An uncontrolled operation in the open and underground mine pits might cause a change in the type of land surface, thereby harming adjacent ecosystems.

Satellite imagery and predictive models allow us to track temporal and spatial changes in vegetation conditions of territories near mine sites. By studying spectral vegetation and bare soil condition indices we can monitor and evaluate the environmental conditions of the land which will help in taking appropriate actions toward sustainable mining while avoiding hazards.

Tell Us About Your Latest Visit To A Romanian Mining Site.

All the GoldenEye project consortium teams met at the General Assembly in Romania on October 10th, 2022, to discuss the latest progress on the project and present findings. The meeting took place at the Technical University of Cluj-Napoca. Our scientific team was represented by the GoldenEye Project Coordinator Nataliia Borotkanych and Scientist Kateryna Sergieieva.

The meeting included two stages. The first day was devoted to the discussion of technologies for satellite and ground-based monitoring of mining areas proposed and implemented by the project participants. We presented technologies for identifying areas of active quarrying using a time series of radar data, assessing the stability of quarry slopes and dumps based on digital terrain models (Digital Surface Model, DSM), building surface temperature and humidity maps, mapping surface water bodies (technology for identifying water bodies was also presented more explicitly at the IEEE IGARSS 2022 conference).

On the second day, we went to the field testing at the Roșia Poieni quarry, the largest copper mine in Romania where we collected the UAV survey data to build RGB ortho mosaic and DSM. Such data is used by consortium members to detect changes in the pit surface, and also validate and evaluate the accuracy of proposed technologies.

How Will The Technologies Developed By EOSDA Be Used After The Project Is Closed?

The GoldenEye AI platform will be released publicly and be available as a software product. The current field trials will demonstrate the efficiency of the platform-exclusive AI algorithms. At the same time, environmental monitoring capabilities will be available on a continuous basis. The data models require extensive data sources as they will be adjusted for each individual location. For example, a typical business case will look like this: a rare Earth metals mine orders mapping and analytics from the GoldenEye platform for two of their locations. Of course, the AI capabilities can potentially be scaled for global coverage, but in the nearest future, it is planned to provide those on demand.

One great advantage of the GoldenEye platform is that it assists in finding solutions for more sustainable mining production practices. So, not only we can conduct analytics and help the mining and processing enterprises increase efficiency in geological exploration, but also help them protect the environment. However, we can’t force our observations on anyone. We won’t be able to single-handedly initiate environmental monitoring practices. Instead, mining and processing enterprises are encouraged to use the platform for their own benefit and the benefit of the Earth.

How Do You See EOSDA’s Assistance With Projects Like Horizon 2020 In The Future?

I can highlight our capabilities in relation to Horizon-like projects in a few industries. In Europe, the company can provide expertise for agri-environmental monitoring by leveraging the proprietary algorithms that are implemented on our SaaS platforms EOSDA Crop Monitoring and EOSDA Forest Monitoring.

The platforms’ functionality is based on satellite observation data, processed by our machine learning models, and enriched by analytical insights that are provided in a user-friendly and actionable web interface (mobile apps are also available for scouting). Our software products and services make it easier to comply with the principles of rational nature management and sustainable development, ensuring food security. Observations and analytics targeted at forestry can help avoid the growing deforestation tendency and track forest fires, as well as other forest health indicators.

Custom projects of our team like mining site monitoring can extend to a variety of different industries that can potentially benefit from satellite imagery and data analytics.

By 2025, we will be able to provide high-resolution data from our EOS SAT constellation of satellites. As of now, we have launched the first one of them into the Earth’s low orbit, and six more will be launched over the next two years. We have a strong team of more than 60 scientists, including 25 Ph.Ds, and we’re currently open to exploring opportunities for collaboration.

EOS Data Analytics has ambitious plans on improving its satellite analytics platforms and is sure to help even more organizations and individuals fight climate change and reach sustainability in their nature conservation efforts in the following year