CROP YIELD ESTIMATION IN THE FIELD LEVEL USING VEGETATION INDICES

Remote sensing can be very useful tool for agriculture management. In this study, remote sensing methods were applied for yield estimation in the field level. There were compared remote sensing data together with yield data obtained from the field. The study area is located in Polkovice in Olomoucký region and a crop planted there in the year 2016 was spring barley as one of most important crops grown in the region. The study area in Polkovice is located at lower elevations with intensive crop production and is climatologically warmer and drier than other areas of the Czech Republic. Year 2016 was the first year when the harvest device has been used for yield analysis in this study area. The output of this method is the yield map displaying the amount of crop harvested in the particular place in the field. The yield data from the field were then compared with remote sensing data in the form of vegetation indices. Two of them were used for comparison – Normalized Difference Vegetation Index (NDVI) and a two-band Enhanced Vegetation Index (EVI2). These indices have been often used for yield estimation in different studies but mostly in larger scales. This study investigates use of NDVI and EVI2 at more detailed scale while using various remote sensing methods. Comparisons show that remote sensing data can provide accurate estimation and can be used for yield forecasting or supplement traditional ways of yield estimation. Results of the study show that yield-index correlations are stronger for satellite data than for the drone data. NDVI showed slightly stronger correlations than EVI2. Strongest correlations between vegetation indices and yields were found for NDVI from Sentinel 2.