Ensuring crop health: Pest detection in arable lands

picture with map zones generated by ai

Arable crops are, undoubtedly, of great economic and social importance for the EU, as half of the total area is used for cereal cultivation. Interestingly, in 2017 the EU produced 309.9 million tonnes of cereals (which is almost 11.9 % of the global harvest) like wheat, maize, barley, rye, and oats, but also rice. However, pest infections in arable crops are quite common, posing a serious threat to agriculture and food security, with major yet unaccounted yield losses. Eden Library Viewer, due to its morphological and operational characteristics, is currently used in orchards and vineyards, as it can easily be attached to any tractor, sprayer, or ATV and detect anomalies in trees and vines using computer vision technologies. But what about arable crops? Is there any alternative to identify pest hotspots early and prevent their spread in such fields?

Conventional methods for real-time pest scouting are insufficient for covering the needs for spatial and temporal resolution data that are required to achieve optimal management of agricultural resources. These methods are usually costly and time-consuming, not to mention that they don’t keep up with the fast increase of pests infestations caused by climate change. As a result, local and global organizations lack the necessary data for large-scale policymaking, farm statistics, and precise intervention strategies. 

Earth Observation (EO) tools together with high-resolution and openly available Sentinel data can be used to overcome the data-scarce environment, assist statistical reporting, monitor plant health over large areas, and prevent diseases at the early growth stages. In fact, the Sentinel-2 mission is characterized by satellite imagery with multi-spectral data from 13 bands in the visible, near-infrared, and short-wave infrared parts of the spectrum. After that, multispectral remote sensing data need to be combined with computation technologies, advanced analytical methods, and machine learning algorithms in order to map the extent of each incursion through geospatial modeling. Later on, the maps of crop stress have to be linked to biophysical data, vegetation indices, and agro-climatic factors favorable for pests. Thus, the drivers of pest diseases can be well-understood and a risk assessment for potential outbreaks can be conducted. By predicting the spatial distribution of important invasive pest populations, we are capable of developing statistics that can be incorporated into national agriculture statistics and accounting systems.

 So, to prevent plant diseases from spreading and to ensure productive and sustainable agriculture, robust and regular monitoring of plant health is essential. Also, information on climate conditions, land uses, soil properties, and farming practices is needed for valuable insights into the dynamics of pests. The resulting data from the above-mentioned data-processing and analysis procedures can serve as a baseline to design targeted intervention measures. As a result, insects, bacteria, viruses, weeds, and fungi will not undermine efforts to increase the productivity of cereals anymore and farmers will be facilitated to minimize the amounts of pesticides they use.