Your territories seen from the sky.
Your decisions grounded in data.

Problem

Roads cut off, no connectivity, entire days of ground teams mapping damage. Resource allocation decisions are made blind.

Drone + AI solution

A drone surveys the affected area, AI analyses 2,000 buildings in 7 minutes, classifies damage into 4 severity levels, and produces a map of passable roads. Reference: CLARKE system (Texas A&M), >90% accuracy.

Problem

No reliable cadastre, impossible to measure cultivated areas, yields remain estimated. Food security programmes lack baseline data.

Drone + AI solution

An eBee X maps 500 ha in a single flight, 3 cm precision. AI automatically segments plots, detects water stress through multispectral imagery.

Problem

Camps evolve too fast, structures are tiny, and ground-based counts are slow and inaccurate. Impossible to plan basic services without reliable data.

Drone + AI solution

Kakuma, Kenya — 102 drone flights, 161,000 images. AI automatically identifies tents, solar panels, and latrines to estimate population and needs.

Problem

REDD+ projects require precise forest cover data. Satellite images are often blocked by clouds in tropical zones, making monitoring impossible.

Drone + AI solution

The drone flies below the clouds, producing orthophotos at 2–5 cm/pixel. AI quantifies deforestation through automated change detection.

Problem

The EU, United Nations, and AFD require evidence-based results. Narrative reports are no longer sufficient.

Drone + AI solution

A drone flight combined with AI produces measurable, georeferenced, dated, and reproducible indicators. Data that directly meets donor requirements.

Context

Comprehensive overview of the drone ecosystem for international cooperation. When is a drone relevant? When is it not? How to integrate a drone component into a donor-funded project?

Sessions

• A.1 Drone typologies and capabilities
• A.2 Drones vs satellite vs field — decision comparison
• A.3 Uses in cooperation (WFP, UNICEF, Flying Labs)
• A.4 Drone project cycle
• A.5 Key ecosystem actors
• A.6 Drones in donor frameworks (INTPA, PRAG, ECHO)

Context

From flight planning to exploitable thematic map. This module covers the entire chain: sensors, photogrammetric processing, derived products, and GIS analysis.

Sessions

• B.1 Mission planning (GSD, overlap, GCP)
• B.2 Sensors and imagery (RGB, multispectral, thermal, LiDAR)
• B.3 Photogrammetric processing — Practical OpenDroneMap/WebODM
• B.4 Derived products (DSM/DTM, 3D model, point cloud)
• B.5 GIS analysis — Practical QGIS
• B.6 From data to decision

Context

The differentiating module. Makes AI accessible to non-IT professionals. From semantic segmentation to object detection, participants learn to leverage AI for decision-making.

Sessions

• C.1 Introduction to AI and deep learning for non-IT professionals
• C.2 Semantic segmentation (U-Net, >95% accuracy) — Practical
• C.3 Object detection (YOLO v12, Mask R-CNN, F1 >90%) — Practical Ultralytics
• C.4 Change detection (automated Before & After) — Practical
• C.5 Vision Language Models (GPT-4V, Claude, Gemma)
• C.6 Reference datasets (xBD: 850K+ annotations, CLARKE, EBD, LADI v2)
• C.7 Integrated practical

Context

Selectable sub-modules based on your organisation's intervention sector. Each sub-module combines real cases, field data, and practical exercises.

Sessions

• D.1 Agriculture and food security
• D.2 Biodiversity and protected areas (WWF CANOPÉE-TRIDOM case)
• D.3 Forests, deforestation, and carbon (REDD+)
• D.4 Urban planning and territorial planning (Caye Sable case)
• D.5 Land governance and participatory cadastre
• D.6 Water resources and coastal erosion
• D.7 Infrastructure and rural roads

Context

The flagship sub-module: post-disaster needs assessment (PDNA). From drone flight to AI-automated damage map, in a real humanitarian context.

Sessions

• E.1 Post-disaster damage assessment — complete pipeline (ref. CLARKE, >90% accuracy, 2,000 buildings in 7 min)
• E.2 Camp mapping (Kakuma Kenya, 102 flights, 161K images)
• E.3 Search and rescue (SAR)
• E.4 Disaster risk reduction (DRR)
• E.5 Humanitarian logistics and delivery (Zipline)
• E.6 Human rights documentation

Context

Navigating regulatory and ethical constraints in each country of intervention. From the EASA framework to African regulations, including Do-No-Harm and data protection.

Sessions

• F.1 Regulatory framework (EASA Europe, Africa ~60% countries with rules, Caribbean, ICAO)
• F.2 Drone logistics on mission (import, insurance, customs)
• F.3 Do-No-Harm (military perception, community consent)
• F.4 Privacy and data protection
• F.5 AI ethics (algorithmic bias, dataset equity)

Context

Hands-on training with real ARKO equipment (eBee X, quadcopters). From mission preparation to the final report, including field flight and integration into a donor logical framework.

Sessions

• G.1 Mission preparation (operational checklist)
• G.2 Real drone flight practice (min 4h field)
• G.3 Processing and analysis of collected data — Practical
• G.4 Drone component design in projects (logical framework, PRAG/RBM/MEAL)
• G.5 Budget, ToR, and indicators
• G.6 Project presentations — debriefing