The reduction of unwanted plant species in pastures is a persistent objective of grassland management. Evaluating different management options requires the assessment of the spatial coverage of the unwanted species. Here, we evaluate the use of drone-based images to quantify the cover of buttercup (Ranunculus acris) in an upland pasture (1654 m asl.) in the Central Swiss Alps. Buttercup is of primary concern because it is moderately toxic and avoided by grazers. Between 2016 and 2020, we conducted a randomized complete block trial with ten different treatments (combinations of grazing, mowing, liming, herbicide and overseeding) in four repetitions. Aerial images were taken annually at the peak of buttercup flowering, with a fixed-wing autonomous drone (senseFly eBee) carrying an RGB camera (Canon S110 and from 2019, senseFly S.O.D.A.) and post-processed using Pix4Dmapper. Yellowness was calculated as the percentage of yellow pixels using optimized thresholds on the RGB channels. The correlation coefficient between the yellowness of the images and the share of buttercup estimated by an independent observer was above 0.85 for the last two years. The newer S.O.D.A. camera outperformed the S110 due to its higher resolution, which was shown to be crucial for this kind of assessments.