COBECORE
The Congo basin eco-climatological data recovery and valorisation (COBECORE) project aimed to establish baseline measurements necessary for long-term (retrospective) ecological and climatological research through the recovery and valorisation of unexplored historical data collected in the Congo Basin by Belgian scientists during the colonial period. The project generated three main data streams through the completion of its four objectives: (1) data recovery of the historical climate record for the central Congo basin; (2) historic metrics of forest structure through digitization of aerial photographs; (3) data recovery of historic leaf traits from herbarium specimens; and (4) data integration and dissemination.
This project (COBECORE, contract BR/175/A3/COBECORE - 584K EURO) was led by BlueGreen Labs partner Koen Hufkens as Principal Investigator in collaboration with partners at various Belgian research institutes and successfully completed in September 2022. The COBECORE project implemented state-of-the-art digitization techniques, including machine learning, citizen science and several European collaborations, resulting in practical insights for future digitization projects, outreach for secondary schools and public interest, and numerous publications in A1 scientific journals. The data recovered during COBECORE continues to inspire new research opportunities and remains a valuable reference for contemporary research.
and beyond…
The project sees steady continuation in the development of handwritten text recognition (HTR) software in collaboration with Prof. dr. Wim Thiery at the Free University Brussels and the Digital Humanities department of Ghent University. Dr. Hufkens also lectures on HTR best practices within the context of data recovery efforts.
Highlighted achievements
- conceptualization of a software framework for automated transcription (weaHTR)
- Digitized all tabulated climate data stored in the Royal Archives (~75K scans)
- Transcribed a subset using a highly successful citizen science project (Jungle Weather)
- Extracted leaf traits using Machine Learning from herbarium specimen
- A spin-off Machine Learning for high-school students (KIKS)
- Recovered and valorized historical remote sensing data combining Structure-from-Motion and Machine Learning.
- All workflows and most data are openly available (if not, upon request due to data volumes)
Select publciations
- Vercruysse et al. 2025 - Human-in-the-loop tabular data extraction methods for historical climate data rescue
- Brönniman et al. 2019 - Unlocking pre-1850 instrumental meteorological records: A global inventory
- Hufkens et al. 2020 - Historical Aerial Surveys Map Long-Term Changes of Forest Cover and Structure in the Central Congo Basin
- Meeus et al. 2020 - From Leaf to Label : A Robust Automated Workflow for Stomata Detection
- Bauters et al. 2020 - Century‐long Apparent Decrease in IWUE with No Evidence of Progressive Nutrient Limitation in African Tropical Forests
Relevant links
- The project website, cobecore.org
- A browsable index of recovered Agricultural Research Archive resources
- The KIKS Machine Learning course for secondary school students using stomata data