INSECT DETECTION FOR HI-TECH FARM PEST SURVEILLANCE
What is the issue?
Insect pests can cost famers huge amounts of money in lost production. Insecticides help control the problem but increasing concerns about environmental damage have led to widespread restrictions on use. Pesticides can also have deleterious effects on beneficial insects. If farmers knew exactly where pests were in their fields and could target hotspots early on, less spraying would be needed. Insect traps are currently used to assess numbers, but these can take several days to indicate an outbreak and they do not pinpoint where in a field the pests may be concentrated.
“If we can better identify the patterns of insect immigration into crops then we only have to treat areas of the crop that are at risk. The partners have been fantastic and very supportive in encouraging us to publish.” Dr Sam Cook, Rothamsted project leader
What did we do?
Drawing on years of entomological expertise, Rothamsted partnered with Fauna Photonics and Xarvio (a subsidiary of BASF). Fauna Photonics have designed a LIDAR system that uses laser light to detect insects in flight. Rothamsted scientists flew thousands of insects of 27 species commonly found on oilseed rape through the detector which recorded the unique flight pattern and shape of each species.
Machine learning was then used to develop sophisticated algorithms to automatically identify one species from another. The system will ultimately be able to automatically detect and identify individual insects in real time.
Impact/expected outcomes If the project is successful, farmers will ultimately be able to receive real-time information on insect species in their crops via mobile phones or tractor mounted computers. That will enable them to quickly identify problematic hotspots and deal with outbreaks before they spread. This will mean less pesticide use and potentially more targeted interventions that avoid damage to beneficial species.