GoldenEye - Press Release

‍

GOLDEN AI - THE EARTH OBSERVATION AND GLOBAL NAVIGATION SATELLITE SYSTEM DATA ACQUISITION & PROCESSING PLATFORM FOR SAFE, SUSTAINABLE & COST-EFFICIENT MINING OPERATIONS

‍

Summary:

OPT/NET played a key role in the GoldenEye HORIZON2020 project, a 3,5-year initiative funded by the European Commission. As a key result, the GOLDEN AI extension module was developed as an upgrade for OPT/NET’s commercial MONITORED AI^TM platform. In this platform, we combine satellite, drone and sensor data, and perform advanced analysis of collected data to enhance mine site safety, conduct environmental monitoring and oversee mine production. The collected data is accessible, actionable and supports informed decision-making.

‍

‍Background:

In mid-2020, OPT/NET B.V., as part of a broader consortium of 17 European companies and research partners, was granted funding by the European Commission to participate in a H2020 project. OPT/NET’s role in the project was to develop an Artificial Intelligence enabled geospatial analytics platform for monitoring and analysis of mine sites across Europe.

OPT/NET's involvement in the GoldenEye project as a technology leader and developer was driven by its extensive experience in providing artificial intelligence solutions for mission-critical data-driven areas.

‍

Project tasks:

The GoldenEye project had the following tasks:

• Develop new sensing technologies and data integration capabilities
• Provide new actionable data for the trial mine sites
• Provide an analytics platform which can fuse the new and existing heterogeneous geospatial data produced in the trials
• Provide a European cloud-based solution for the mining end users and authorities for interpretation the  extracted information

Primarily,  OPT/NET aimed to increase the use of Earth Observation Data produced by Copernicus missions, in combination with extremely high-resolution  UAV produced data, as well as data from novel proximal sensors and manual in-situ measurements, in the scope of the mining industry.

‍

During the GoldenEye project, “GOLDEN A”l was developed as an extension for the OPT/NET’s flagship product line - MONITORED AI^TM, which allows high-resolution data collection, processing, and conversion of raw inputs data into analytical insights. The main objectives of GOLDEN AI are to improve mine safety, environmental and geohazard monitoring, operational management, mineral detection, and production output increase for five pilot sites in Europe.

‍

End users are provided with visualised intelligent maps at their disposal, offering comprehensible information and insights. By using the AI platform, custom solutions and applications can be built that fulfil specific user needs across diverse mining domains leading to substantial and proven time and cost savings, increased efficiency and productivity in the referenceable pilot test cases concluded during the project. 

‍

With the developed GOLDEN AI mining modules for MONITORED AI^TM platform, the repetitive and routine tasks covering the selected mining industry use-cases can be automated, freeing up human analysts and operational resources to focus on more effective, creative, and complex analysis and data driven decision making. These advantages collectively contribute to the transformation and enhancement of mining industry processes. 

‍

Methodology:

GOLDEN AI is built within the framework of the commercially available MONITORED AI^TM product line, developed by OPT/NET over many years, supported with European and local government grants and both internal and external investments. The platform advancements are achieved by using OPT/NET's innovative modular AI methodology, which allows tailoring of the platform to various industries by “plugging in'' relevant technology, task specific modules and AI Knowledge Packs (data processing “cook books” for the AI). Besides the main modules described further in the text, additional modules for 3D underground and above ground object visualisation, with the possibility of terrain extrusion and helicopter mode, have been introduced to enhance the platform’s usability via the sister Horizon project AGEMERA. 

‍

To further advance the added value in the mining domain, AIKPs developed by the experts during the duration of the GoldenEye project resulted in functionalities such as change detection, terrain classification, surface displacement or soil moisture mapping and minimised the use of operational resources at the mines, both in personnel and materials, by shortening the time and reducing the environmentally harmful blasting operations.

‍

Project realisation :

‍

‍

The platform consists of 3 main modules:

•  OCLI, the platform's backend module, provides a pipeline that processes AI Knowledge Packs (AIKPs) and handles the data mining process. It involves automated functions for data acquisition, data preprocessing, image processing (e.g., denoising, edge detection, etc.), and AI processing. In particular, for a given area of interest, it first collects the relevant satellite data products from either DIAS services, Euro Data Cube (EDC), EOS or Supplier’s API. Finally, it prepares analysis ready data (ARD) by performing data cleaning, transformation and filtering. 
• Golden AI GUI, the platform’s frontend module, acts as a web-based general data repository and visualisation platform in Cloud Optimised GeoTIFF (COG) format powered by OGC web services. The data can be interactively explored by the general public and mine site owners, with different authentication and authorization access levels, in an easy, quick and efficient way, assisted by a conversational AI agent.
• Data Repository, which consists of two components: bulk data warehouse and metadata database. Amazon S3 Storage Service is selected for primary storage as a service that provides object storage through a web service interface.

‍

‍

With the GoldenEye project coming to its end in October 2023, an impressive amount of Earth Observation Data has been stored in the platform’s Database and is accessible via the Graphical User Interface, with around 800 data entries, of which almost half cover underground GNSS and geomorphologic remote sensing field trials in Pyhäsalmi (Finland).

‍

Regarding satellite imagery, the most commonly used data used were optical satellite data sources like Sentinel-2, ASTER, Landsat-8, all of which are open-source, and WorldView-3, a commercial satellite data acquisition. It’s worth noting that almost 20% of Earth observation Data came from radar imagery provided by Sentinel-1, a part of ESA’s Copernicus mission, and almost 6% of detailed optical imagery collected from drone flights.  

‍

Conclusion:

GOLDEN AI showcases the possibilities of MONITORED AI^TM, OPT/NET’s product line that utilises the modular structure backed up by powerful Artificial Intelligence Knowledge Packs capable of complex AI analysis tasks. The platform itself encompasses everything from satellite data acquisition, data fusion, data preprocessing, AI analytics and finally data visualisation in the platform's web-based Graphical User Interface. 

The platform combines the processing & automation capabilities of AI with the natural problem-solving abilities of humans, resulting in an interactive, user-friendly and extendable web-based platform for a wide range of industry-specific challenges.

‍

Now, with the end of the project close at hand, the platform holds around 800 data products for five use cases, ranging from mine safety and mineral prospecting to mine post closure, which allows site owners, stakeholders and the general public to interactively explore and utilise the data for making informed data-driven decisions. 

‍