Blog – Analysis and expert opinion
5 MIN. READ

Résultats du Projet GEMA – Génération de Cartes à Haute Résolution et Interprétation Avancée de Données Géologiques grâce à l’Intelligence Artificielle

Au cours de ses 18 mois d’exécution (août 2023 – février 2025), le projet GEMA a réalisé des avancées significatives dans l’intégration de l’intelligence artificielle appliquée à la géophysique, atteignant des jalons techniques qui apportent de la valeur à la fois à la recherche et à l’exploration appliquée.

Proyecto GEMA

Développement d’une base de données internationale
Comme point de départ, une base de données géophysique internationale a été construite, intégrant des données radiométriques, magnétiques, hyperspectrales, satellitaires et des modèles numériques de terrain. Cette collecte de données, obtenue dans différentes régions et sous diverses conditions opérationnelles, a permis de disposer d’un référentiel robuste pour l’entraînement et la validation des modèles d’intelligence artificielle, garantissant leur capacité de généralisation.

Méthodologies avancées de traitement

  • Le projet a mis en place un pipeline reproductible de traitement et de normalisation des données, incluant l’homogénéisation des systèmes de référence spatiale (CRS) et la correction des artefacts fréquents dans les données magnétiques. Ces méthodologies ont réduit de manière substantielle les temps de préparation des données, tout en améliorant la qualité et la traçabilité des informations utilisées dans la modélisation.

Modèles d’intelligence artificielle développés

  1. Correction des artefacts magnétiques (Levelling) : Une approche basée sur des modèles bayésiens et l’interpolation spatiale a été mise en œuvre, donnant naissance à un algorithme capable de réduire significativement la nécessité d’ajustements manuels. Cet outil, également déployé sous forme de plugin, a démontré son utilité pour homogénéiser les levés aérogéophysiques de manière plus rapide et cohérente.
  2. Segmentation automatique des linéaments : Des réseaux neuronaux convolutifs (CNN, architecture U-Net) ont été entraînés pour détecter et tracer des structures géologiques sur des cartes magnétiques. Les résultats ont atteint des valeurs de précision élevées (avec un coefficient de Dice proche de 0,8 sur les ensembles d’entraînement), démontrant la capacité de ces techniques à automatiser des tâches qui nécessitaient traditionnellement des mois d’interprétation manuelle.

Prévision géomagnétique pour le forage directionnel (IFR2) : Un modèle spatio-temporel (INLA-SPDE) a été développé pour prédire avec une haute résolution les variations du champ magnétique dans les zones de forage. Cela a permis d’optimiser l’orientation des forages et de réduire les erreurs de trajectoire, avec une validation effectuée même lors d’événements géomagnétiques spécifiques observés durant la campagne de validation.

Validation et résultats quantitatifs
Les progrès ont été vérifiés par des processus de validation rigoureux :

  • Les algorithmes de levelling ont été comparés à des données de vols corrigés manuellement, montrant une forte cohérence spatiale et une amélioration des coefficients de corrélation (R²).
  • Les modèles de linéaments ont été évalués avec des métriques standards (précision, rappel, F1 et Dice), confirmant leur capacité de généralisation dans des régions non incluses dans l’entraînement.
  • Le modèle de prévision géomagnétique a été validé par comparaison minute par minute des composantes Fx, Fy et Fz avec des stations indépendantes, atteignant de faibles erreurs absolues moyennes (MAE) et anticipant avec succès des épisodes de variabilité magnétique.

Portée géographique de validation
Les méthodologies ont été testées dans divers contextes géologiques internationaux, renforçant leur polyvalence.

Les zones pilotes incluent :

  • Afrique : Ouganda, République Démocratique du Congo (Katanga), Nigéria.
  • Europe : France, Finlande et Russie.
  • Asie : Kazakhstan, Mongolie et Turquie.
  • Amérique du Nord : Canada.

Cette portée géographique illustre l’ambition globale du projet et garantit que les outils développés soient applicables dans différents contextes géologiques.

Financement
Le projet GEMA dispose d’un budget total de 1.418.908 €, dont 1.206.071,80 € correspondent au soutien de l’Union européenne (Subvention Brute Équivalente). Ce soutien a été essentiel pour le développement de solutions innovantes en intelligence artificielle et leur application à l’exploration géophysique.

Related news

Contact us.

FILL THE FORM
Blog – Analysis and expert opinion
5 MIN. READ
4th International Conference on Financing for Development

Public-Private solutions in the spotlight

  • From 30 June to 3 July 2025, Seville hosted a historic gathering: the Fourth International Conference on Financing for Development (FfD4). A decade after the last edition in Addis Ababa, this long-awaited event brought together over 180 national delegations, including nearly 70 Heads of State, to shape the future of development finance.
    Xcalibur Smart Mapping was proud to attend as an accredited participant, with a delegation composed of Andrés Blanco, Andrés Niño, Víctor González, and Paola Lancellotti. The team was actively involved throughout the week, engaging in more than 20 side events—with a particular focus on the International Business Forum, which elevated the role of private sector actors in achieving global development goals.

High-Level Engagements & New Connections

Our team had the opportunity to meet with key stakeholders across public and multilateral institutions, including:
✔️ Ministries of Egypt and Zimbabwe
✔️ President of the European Investment Bank (EIB)
✔️ Representatives of the OPEC Fund, Islamic Development Bank, and Asian Infrastructure Investment Bank
✔️ Spanish institutions and development finance bodies
These engagements helped position Xcalibur’s unique offering—non-invasive airborne geophysics and pre-competitive geoscientific data—as a vital tool for natural resource governance, infrastructure planning, and climate-smart investment.

Why FfD4 mattered to us

FfD4 made it clear: ODA alone isn’t enough. With an estimated need of $4.5 trillions to meet the Sustainable Development Goals (SDGs), innovative public-private solutions are now essential.
Among the most relevant developments:
As a private sector leader in geoscientific innovation, Xcalibur’s work directly supports de-risking investment and enabling sustainable development, especially in emerging economies rich in natural resources.
A Milestone Hosted by Spain
Spain’s leadership under the pillars of Refuge, Reform, and Reinforcement demonstrated a renewed commitment to inclusive development and the mobilization of both public and private actors.
The strong involvement of the business sector—particularly through the International Business Forum—signaled a shift from development as aid to development as investment, innovation, and long-term partnership.
The conference concluded with the adoption of the Seville Commitment for Action
Spain’s effort in hosting this high-level event was widely recognized and reaffirmed its growing role in global development diplomacy.
We return from Seville with renewed energy and a strengthened global network. Our next steps:
– Following up with institutional and multilateral contacts
– Exploring involvement in new development finance initiatives
– Continuing to provide data-driven solutions for resilience, sustainability, and investment readiness worldwide .

We thank all our partners, old and new, and look forward to working together to turn the vision of FfD4 into action on the ground.

Related news

Contact us.

FILL THE FORM

Blog – Analysis and expert opinion
5 MIN. READ

The demand for critical minerals is expected to quadruple by 2040

  • The informative dossier, developed by Xcalibur Smart Mapping based on official sources, shows how the application of artificial intelligence can manage natural resources, anticipate risks, minimize impacts, and build resilience in a transitioning energy system.
  • The application of AI in this field becomes a strategic approach to anticipating the future and better managing the present of the energy transition.
  • 800 billion dollars are needed in the critical minerals sector to reach the net-zero emissions target by 2040.

The demand for critical minerals will quadruple by 2040. This is one of the key findings of the First Xcalibur Smart Mapping Informative Dossier, titled “Artificial intelligence, critical minerals, and sustainable transformation”, developed by Xcalibur Smart Mapping, the global leader in natural capital mapping solutions. Drawing on its expertise and a range of official sources, the report highlights the geological, energy, and technological challenges facing the industry, as well as the transformative role that artificial intelligence (AI) plays in the current energy transition landscape.

The dossier notes that the growing demand for critical minerals is being driven by the development of key technologies such as batteries, electric transportation, and solar energy. To meet the net-zero emissions target by 2040, an investment of 800 billion dollars in the critical minerals sector will be required, according to the International Energy Agency.

The document presents artificial intelligence as a strategic tool for the energy transition. Its application is not only intended to optimize processes but also to transform the way we explore, plan, and manage the planet’s natural resources. AI enables us to anticipate future changes and manage present-day resources more efficiently. In fact, this key enabler extends beyond the energy sector: the dossier highlights that, according to EY’s AI Pulse Survey, 97% of companies that invested in AI in 2024 reported positive returns.

“It is not about technology replacing scientific knowledge, but rather enhancing and scaling it up. Technological integration does not replace experience, it multiplies it. It allows for the development of strategic solutions that improve exploration efficiency, reduce environmental impact, and reinforce traceability. It accelerates the identification and valuation of essential resources—key to European autonomy and sustainability—and transforms data into strategic decisions for a fairer and more resilient future,” highlights Jorge Urios, R&D Director of Xcalibur Smart Mapping.

Transforming geological exploration with AI

The application of artificial intelligence in geological exploration marks a turning point in the way we locate and assess subsurface resources. Unlike conventional methods, which rely on manual interpretation, long lead times, and high uncertainty, new approaches integrate advanced sensors, machine learning, and predictive models that transform millions of data points into actionable knowledge.
AI can identify patterns invisible to the human eye, anticipate risks, optimize resources, and accelerate real-time decision-making. Its implementation improves efficiency and reduces environmental impact by eliminating invasive drilling and prolonged campaigns.

With the need for critical minerals projected in the coming decades, the dossier prepared by the global leader in natural capital mapping solutions underscores the need to adopt technologies that turn data into strategic decisions.

As a result, both work times and financial costs are drastically reduced.
For example, the company’s AI-based tools for aeromagnetic leveling would reduce 150 workdays per year across 30 projects. Meanwhile, the algorithms for magnetic compensation would reduce 80 flights per year, equivalent to approximately 100 tons of CO₂ emissions and a saving of more than $500,000.

“Artificial intelligence and applied knowledge position Xcalibur Smart Mapping as a key player in ecosystem monitoring, territorial management, and environmental traceability. Our solutions go beyond mining: they directly contribute to sustainability and climate action policies,” says Nicolas Leiva, Head of AI at Xcalibur Smart Mapping.

The dossier prepared by Xcalibur Smart Mapping concludes that AI is not just a set of algorithms, but a strategic tool for guiding decisions with a real impact on the planet’s sustainability. Thanks to these algorithms, it is possible to better explore, prioritize areas with a smaller environmental footprint, reduce improvisation, and increase transparency for public and private entities, investors, and communities. Furthermore, the data obtained allows for better management of natural resources, the design of more effective public policies, and the guidance of responsible investments.

For this reason, Xcalibur Smart Mapping has taken on a relevant role in the energy transformation with an approach based on AI, advanced sensors and predictive models that not only respond to the demand for efficiency, but also incorporate a systemic vision: exploring better, with more information, less impact and greater purpose.

If you would like to access the full document, please click the button below.

Related news

Contact us.

FILL THE FORM

Blog – Analysis and expert opinion
5 MIN. READ

Resultados do Projeto GEMA – Geração de Mapas de Alta Resolução e Interpretação Avançada de Dados Geológicos mediante Inteligência Artificial

Ao longo dos seus 18 meses de execução (agosto de 2023 – fevereiro de 2025), o projeto GEMA alcançou avanços significativos na integração da inteligência artificial aplicada à geofísica, atingindo marcos técnicos que agregam valor tanto à investigação como à exploração aplicada.

Proyecto GEMA

Desenvolvimento de uma base de dados internacional
Como ponto de partida, foi construída uma base de dados geofísica de alcance internacional, que integra informações radiométricas, magnéticas, hiperespectrais, imagens satelitais e modelos digitais de terreno. Essa coleta de dados, obtida em diferentes geografias e sob diversas condições operacionais, permitiu dispor de um repositório robusto para treino e validação de modelos de inteligência artificial, assegurando a capacidade de generalização das soluções desenvolvidas.

Metodologias avançadas de processamento
O projeto estabeleceu um pipeline reprodutível de processamento e normalização de dados, incluindo a homogeneização de sistemas de referência espacial (CRS) e a correção de artefatos comuns em dados magnéticos. Essas metodologias reduziram substancialmente os tempos de preparação de dados, melhorando a qualidade e a rastreabilidade das informações utilizadas na modelagem.

Modelos de inteligência artificial desenvolvidos

  1. Correção de artefatos magnéticos (Levelling): Implementou-se uma abordagem baseada em modelos bayesianos e em interpolação espacial, resultando num algoritmo capaz de reduzir de forma significativa a necessidade de ajustes manuais. Esta ferramenta, também disponibilizada como plugin, demonstrou sua utilidade para homogeneizar levantamentos aerogeofísicos de maneira mais rápida e consistente.
  2. Segmentação automática de lineamentos: Treinaram-se redes neurais convolucionais (CNN, arquitetura U-Net) para a deteção e o traçado de estruturas geológicas em mapas magnéticos. Os resultados alcançaram valores de precisão elevados (com coeficiente Dice próximo de 0,8 nos conjuntos de treino), demonstrando a capacidade dessas técnicas para automatizar tarefas que tradicionalmente exigiam meses de interpretação manual.
  • Previsão geomagnética para perfuração direcional (IFR2): Foi desenvolvido um modelo espaço-temporal (INLA-SPDE) capaz de prever, com alta resolução, as variações do campo magnético em zonas de perfuração. Isso permitiu otimizar a orientação de poços e reduzir erros de trajetória, sendo validado inclusive frente a eventos geomagnéticos específicos registados durante a campanha de validação.

Validação e resultados quantitativos
Os avanços foram confirmados por meio de processos de validação rigorosos:

  • Os algoritmos de levelling foram comparados com dados de voos corrigidos manualmente, mostrando alta consistência espacial e melhoria nos coeficientes de correlação (R²).
  • Os modelos de lineamentos foram avaliados com métricas padrão (precisão, recall, F1 e Dice), confirmando sua capacidade de generalização em regiões não incluídas no treino.
  • O modelo de previsão geomagnética foi validado através da comparação minuto a minuto dos componentes Fx, Fy e Fz em relação a estações independentes, alcançando baixos erros médios absolutos (MAE) e antecipando com sucesso episódios de variabilidade magnética.

Âmbito geográfico de validação
As metodologias foram testadas em diversos cenários geológicos internacionais, reforçando a sua versatilidade. Entre as zonas piloto destacam-se:

  • África: Uganda, República Democrática do Congo (Katanga), Nigéria.
  • Europa: França, Finlândia e Rússia.
  • Ásia: Cazaquistão, Mongólia e Turquia.
  • América do Norte: Canadá.

Esse alcance geográfico evidencia a ambição global do projeto e garante que as ferramentas desenvolvidas sejam aplicáveis em diferentes contextos geológicos.

Financiamento
O projeto GEMA conta com um orçamento total de €1.418.908, dos quais €1.206.071,80 correspondem ao apoio da União Europeia (Subvenção Bruta Equivalente). Esse apoio foi fundamental para o desenvolvimento de soluções inovadoras em inteligência artificial e sua aplicação na exploração geofísica.

Related news

Contact us.

FILL THE FORM
Blog – Analysis and expert opinion
5 MIN. READ

Results of Project GEMA – Generation of High-Resolution Maps and Advanced Interpretation of Geological Data through Artificial Intelligence

Over its 18 months of execution (August 2023 – February 2025), the GEMA project has achieved significant progress in the integration of artificial intelligence applied to geophysics, reaching technical milestones that bring value to both research and applied exploration.

Proyecto GEMA

    • Development of an international database
      As a starting point, an international geophysical database was built, integrating radiometric, magnetic, hyperspectral, satellite, and digital terrain model data. This data collection, obtained from different geographies and under diverse operational conditions, has made it possible to create a robust repository for training and validating artificial intelligence models, ensuring the generalization capacity of the solutions developed.
    • Advanced processing methodologies
      The project has established a reproducible pipeline for data processing and normalization, including the homogenization of spatial reference systems (CRS) and the correction of common artifacts in magnetic data. These methodologies have substantially reduced data preparation times, improving the quality and traceability of the information used in modeling.
    • Artificial intelligence models developed
    1. Magnetic artifact correction (Levelling): An approach based on Bayesian models and spatial interpolation was implemented, resulting in an algorithm that significantly reduces the need for manual adjustments. This tool, also deployed as a plugin, has proven useful for homogenizing airborne geophysical surveys more quickly and consistently.
    2. Automatic lineament segmentation: Convolutional neural networks (CNN, U-Net architecture) were trained to detect and trace geological structures in magnetic maps. The results achieved high accuracy (with a Dice coefficient close to 0.8 in training sets), demonstrating the ability of these techniques to automate tasks that traditionally required months of manual interpretation.
    • Geomagnetic prediction for directional drilling (IFR2): A spatio-temporal model (INLA-SPDE) was developed to predict high-resolution variations of the magnetic field in drilling zones. This has optimized borehole orientation and reduced trajectory errors, being validated even against specific geomagnetic events recorded during the validation campaign.

    Validation and quantitative results
    Progress was verified through rigorous validation processes:

    • Levelling algorithms were compared against manually corrected flight data, showing high spatial consistency and improved correlation coefficients (R²).
    • Lineament models were evaluated with standard metrics (precision, recall, F1, and Dice), confirming their ability to generalize in regions not included in the training.
    • The geomagnetic prediction model was validated by minute-by-minute comparison of Fx, Fy, and Fz components against independent stations, achieving low mean absolute errors (MAE) and successfully anticipating episodes of magnetic variability.

    Geographical scope of validation
    The methodologies were tested in diverse international geological scenarios, reinforcing their versatility. Pilot areas included:

    • Africa: Uganda, Democratic Republic of the Congo (Katanga), Nigeria.
    • Europe: France, Finland, and Russia.
    • Asia: Kazakhstan, Mongolia, and Turkey.
    • North America: Canada.

    This geographical reach demonstrates the global ambition of the project and ensures that the developed tools are applicable in different geological contexts.

    Funding
    The GEMA project has a total budget of €1,418,908, of which €1,206,071.80 corresponds to the contribution of the European Union (Gross Equivalent Subsidy). This support has been essential for the development of innovative artificial intelligence solutions and their application in geophysical exploration.

 

Related news

Contact us.

FILL THE FORM
Blog – Analysis and expert opinion
2 MIN. READ

 

 

Related news

Contact us.

FILL THE FORM