Integrating Flir One Pro With Raspberry Pi for YOLOv7 Worm and Egg Detection

Clients

Techstack

Purpose

Spotting worm and egg infestations by eye is slow, easy to miss, and usually catches a problem only after it has spread. A research team wanted to prove that pairing a thermal camera with AI could pick pests out automatically — from heat signatures a human inspector would never notice.

Description

We built a compact device that watches a live thermal feed and runs it through a custom-trained vision model that locates and classifies worms and eggs on the fly. Detections are drawn straight onto the camera stream in a simple monitoring interface, and the model was checked against a labeled set of test images to confirm it held up in practice. It is exactly the kind of embedded computer-vision problem — heat imaging, live inference, low-power hardware — that we take on and ship.