At NBR, were looking to improve our research output by integrating remote sensing, and in particular drones, into our data collection schedule and in the analysis. We are still investigating exactly what information is useful, but the clearest use is to get an overview of all research sites as they progress through the year.

The implementation of these little machines has actually taken up a lot more time than expected. Much of this has been in working on how to integrate the spatial data in our project and data management system. All of it has been enjoyable. I’ve also written one article for Betodlaren and two other blog posts on this (see right/ bottom of this page).


The stable of drones we use in research is likely to continue to grown and evolve with time, and as changes in the law require updating of machines. Better cameras and multispectral are the foreseeable next steps.

Further down the road, I think there is a chance that we’ll come back down to earth. They go by a few names – Unmanned Ground Vehicles (UGVs), Phenomobiles, research-robots – and they often look like equipment that we already see around – the Speedy sprayer, and Farmdroid come to mind – but they work on similar analytics as drones. Why I think that UGVs will be more popular in the future is that they seem to be able to achieve a higher rate of accuracy than drones (like in this cercospora example) owing to their ability to better calibrate, the higher resolution, the different angles, and even the ability to control sampling conditions. This means that they would be able to replace even more of the human-done phenotyping than drones are able to. At the same time, they have a great potential to operate even more autonomously for much longer periods. The biggest roadblocks at the minute is the marginal cost – to get not a whole lot more, they cost a whole lot more. Still, I wouldn’t surprised if we have a few UGVs roaming around our research plots within a decade once the price drops a little.