Theoretical and practical aspects of using satellite observations for crop area estimation – Dr. Sergii Skakun

Crop area estimation is a critical component of agriculture monitoring and, specifically, crop production estimation and prediction. Satellite-based geoinformation products, e.g., crop-type maps, can be used efficiently for crop monitoring. For example, they can be used to define an area frame as a stratifying tool to improve the efficiency of ground measurements. In some cases, they can be used to substitute ground measurements for inaccessible areas (imputation), effectively becoming the only source of information. A typical approach for crop area estimation is when satellite-based maps, either historical or in-season, are used for stratification for subsequent ground data collection or photo-interpretation. Several aspects must be considered for crop area estimation, namely the target crops, associated costs, and the desired precision of the estimates. In this talk, I will provide a theoretical framework for crop area estimation from a stratified random sample when strata are from satellite-based crop maps, how the map quality impacts the efficiency of area estimation, and show practical examples of crop area estimation in Ukraine.

 

During the meeting, modern approaches to satellite image analysis, accuracy assessment of maps, and methods for correcting errors in automated data processing were discussed. Special attention was given to the use of satellite technologies for monitoring crops in Ukraine, particularly in temporarily occupied territories. The discussion also covered how machine learning and artificial intelligence can enhance accuracy and contribute to addressing key agricultural and environmental challenges.