AI-vision for integrated management of pest insects and their natural enemies
Ritter Guimapi · AI for Biodiversity Data
Summaries and key moments are machine-produced. Check the recording for the speaker’s exact words.
The short versionMoving precision spraying from static weeds to mobile pests and their natural enemies is blocked mainly by the difficulty of collecting representative image data.
What this was about
Ritter Guimapi of NIBIO presented an idea and work in progress: extending precision sprayer robots, which in Norway have cut herbicide use by 40-90% by dropping herbicide only on detected weeds, to insect pests within conservation biological control. The vision is a robot that distinguishes pests from beneficial natural enemies (e.g. parasitoids) and applies insecticide only to pests. He described why collecting representative training images has proven much harder than for weeds: pest-enemy interactions are transient and hard to time, insects are mobile, damaging life stages sit on leaf undersides, affordable equipment is needed, and regulations restrict releasing natural enemies in controlled settings. He invited suggestions on equipment and approaches.
Why it matters. Integrating AI vision into conservation biological control could reduce pesticide use while protecting beneficial insects, but data acquisition is the bottleneck.
In the room
- Integrated Pest Management and conservation biological control aim to keep pests below thresholds while supporting beneficial organisms.
- Precision sprayer robots in Norway reduce herbicide use by 40-90%, helped by regulations banning drone spraying.
- Transfer learning was the first idea, but field data collection failed.
- Challenges: timing of interactions, insect mobility, eggs and larvae on leaf undersides, cost of equipment, and national regulations.
- A controlled environment is being explored but is not simple either.
Notable moments
Transcript
Automatically generated captions can contain mistakes, especially in names and technical terms. Times are relative to the room recording.
Read the transcript ↓
Related talks.
Scalable Edge AI for Real-Time Biodiversity Monitoring: tracking invasive plant species in roadside imagery
Distilling a tiling-based plant identifier into a single ConvNeXt model makes high-resolution roadside invasive-species monitoring fast, cheap and more accurate.
From Standardized Field Methods to AI-Ready Knowledge: The LEPMON Project with the LAUP Infrastructure Stack for Automated Moth Monitoring
LEPMON's LAUP portal shows an end-to-end infrastructure for camera-based moth monitoring that builds Camtrap DP into routine data management, making millions of observations downloadable and analysable.
Agentic AI Systems for Specimen Interaction and Camera-Based Observation
Off-the-shelf local vision language models can pick out and count specimen morphology and spot damage with minimal guidance.