Earth Observation Project

Earth Observation project – Crops classification in multi-crop farm using satellite images

The National Institute of Statistics of Rwanda (NISR) conducts Seasonal Agricultural Survey since 2013 in order to provide timely, accurate, reliable and comprehensive agricultural statistics in Rwanda in terms of land use, crop production and livestock to monitor current agricultural and food supply conditions and to facilitate evidence based on decision making for the development of agriculture sector. However, it has been a challenge to automatically identify and quantify individual crops in multi-crop farms to estimate production of major crops (MT), Yield in Kg/ha, cultivated area. Besides, the existing methods of crop identification are manual. Consequently, the Data science Team has started a research project on analyzing individual crops in mixed crops farms using Machine Learning techniques on the existing and new high resolution images datasets.

Planned Outputs

• Crops classification in mixed crops farms
• Estimation of Individual crops quantity
• Individual crops cultivated area computation
• Automating individual Crops’ statistics production
• Capacity building of NISR employees