Phenological analysis of unmanned aerial vehicle based time series of barley imagery with high temporal resolution

Abstract

Emerging strategies and technologies in agriculture, such as precision farming and phenotyping depend on detailed data on all stages of crop development. Unmanned aerial vehicles promise to deliver such time series as they allow very frequent measurements. In this study, we analyse a field trial with two barley cultivars and two contrasting sowing densities in a random plot design over 2 consecutive years using the aerial images of 28 flight campaigns, providing a very high temporal resolution. From empirically corrected RGB images, we calculated the green-red-vegetation-index (GRVI) and evaluated the time-series for its potential to track the seasonal development of the crop. The time series shows a distinct pattern during crop development that reflected the different developmental stages from germination to harvest. The simultaneous comparison to ground based assessment of phenological stages, allowed us to relate features of the airborne time series to actual events in plant growth and development. The measured GRVI values range from −0.10 (bare soil) to 0.20 (fully developed crop) and show a clear drop at time of ear pushing and ripening. Lower sowing densities were identified by smaller GRVI values during the vegetative growth phase. Additionally, we could show that the time of corn filling was strongly fixed and happened around 62 days after seeding in both years and under both density treatments. This case study provides a proof-of-concept evaluation how RGB data can be utilized to provide quantitative data in crop management and precision agriculture.