News Agriculture, Climate

Remote mapping by using unmanned aerial vehicles, i.e. drones, brings the measurement of the quantity and quality of grasslands into this millennium. This fast and accurate method produces significant benefits for agricultural entrepreneurs, but it also needs more advanced artificial intelligence solutions alongside it. New solutions are being developed in research projects and at the DroneOlympics event held in Finland in September.

Key quality factors regarding green fodder include its digestibility and nitrogen content. For example, the growth rate has an impact on digestibility, which means that the feed value of green fodder depends significantly on correctly timed harvesting. In addition, the nitrogen content and the harvest volume serve to interpret the amount of nutrients consumed by the crops and the amount of fertilisers needed for the next harvest. When carrying out measurements, speed is the key, as the digestibility may deteriorate significantly even during a single day.

Cooperation projects between the Finnish Geospatial Research Institute (FGI) of the National Land Survey of Finland and the Natural Resources Institute Finland (Luke) have shown that drone sensors and machine learning can help to measure the quality and quantity of grasslands not only quickly, but also accurately. Luke has produced grasslands with different properties for research purposes.

Financial benefits for farmers

FGI’s DroneFinland team has developed tools for sending measurement data directly to tractors operating in the fields, even in real time. Oiva Niemeläinen, a senior scientist at Luke, describes the results of the cooperation projects as a significant breakthrough. Niemeläinen has been in charge of the planning and practical arrangements of grassland testing.

“Measuring grassland yields is a laborious process, which is why parcel-specific yields per hectare are not very well known. Evaluating the quantity and quality of grasslands right before harvesting is a major step forward,” Niemeläinen says.

“By using these new methods, results can be measured directly in a digital format, even for every square metre, if needed. If, as these test results indicate, the mapping results can be reliably generalised over large field areas, farmers will have access to important supporting data regarding the timing of harvesting and grassland fertilisation during the growing season. This, of course, also produces financial benefits.”

However, there is still a lot of work to be done in order to ensure that the research results can be generalised.

“Grasslands consist of diverse mixed crops, and mapping is carried out in varying weather conditions. We need to conduct more research to turn the mapping results produced in different conditions into interpretations of the quantity and quality of crops before these methods can be applied in practice to grassland cultivation,” Niemeläinen says.

Artificial intelligence to help in decision-making processes

Measurement data alone is not enough – it needs to be processed so that it offers help in decision-making processes.  Drones would make the work of farmers easier.

“We need diverse AI solutions. Machines need to understand how, where and when to operate after data about quantity and quality has been measured. By using AI developed on the basis of conditions defined by farmers, we can apply these steps we have now taken in the use of drones to practical work on farms,” says Jere Kaivosoja, research scientist at Luke.

The use of drones will also be developed at the DroneOlympics at the beginning of September. Luke and FGI will hold two innovation challenges at the event: Flowering Forest and Count the Grass. More information about these challenges is available on the DroneOlympics website.

See also