Mini-Grant Award: A Machine Learning Approach for Estimating Maize Yield Using Bio-Physical Parameters Retrieved from Remote Sensed Dataset

Crop yield prediction early in the planting is a crucial facet for decision makers both at the regional and national levels. It helps decision makers to plan based on the projected yield. If high yield is expected, decisions are made on how to manage excess produce and strategies to shield farmers from losses are instituted. Likewise, if poor yield is predicted immediate contingencies measures are instituted to ensure food security.

Maize is the stable food of Kenya and estimates of the expected yield is crucial to aid government on making decisions on measures it should institute to ensure food security in the country.

This project proposes to exploit remote sensed data and machine learning to develop a maize yield prediction model for small scale farmers. Satellite- based optical and radar images have been exploited extensively for monitoring agricultural parameters. An integrated system constituting of machine learning and satellite technologies can be used in forecasting of crops’ yield and drought monitoring and assessment.

Major advantages of crop yield forecasting before harvesting are decisions of agricultural policy, Future crop planning and diversification, Drought declaration and contingency planning during shortage in food grain and Support in crop damage-assessment.


  1. Dr Dennis Kaburu, Jomo Kenyatta University of Agriculture and Technology (JKUAT)
  2. Dr Peter Ochieng, Taita Taveta University
  3. Dr Fridah Kirimi, JKUAT

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top