Barley yield estimation with sentinel-2 vegetation indices

Food and agriculture sector forms the largest piece of gross world product. Due to the increasing world population, global climate change, environmental deteriorations such as nitrogen pollution and desertification, agriculture sector is expected to face significant problems to supply the drastically increasing demand. Timely and accurate estimation of agricultural crop yields is essential for food security; in addition strategically and economically significant for a wide range of actors on the sector; from singular producers to governments. Remote sensing provides more practical and less costly solutions than conventional methods for crop yield forecasting. Sentinel-2 satellites have a great potential for agricultural remote sensing applications due to its vegetation bands and also having the highest spatial resolution in red, green, blue and NIR bands within the satellites whose data is open and free publicly. In this work, the use of Sentinel-2 vegetation indices for yield estimation of barley is analyzed. Commonly used vegetation indices are calculated from two closely dated Sentinel-2 images acquired before the harvest. Linear regression models are constructed between the indices and actual yields. Correlation coefficients are found promising for the yield estimation of barley. Highest correlation coefficients are obtained by the Modified Simple Ratio index.