
Lecturer: Borja Espejo-Garcia
Duration: 2 hours
Participants: 10 people
This workshop introduces key object detection models used in agricultural applications, focusing on Faster R-CNN and YOLO. Participants explore the differences between two-stage and single-stage detectors, comparing trade-offs in accuracy and speed.
The session includes practical labs on fine-tuning a YOLO model for weed detection, data augmentation techniques, and building core object detection components from scratch. Attendees gain hands-on experience with dataset loading, bounding box annotation, training configuration, and performance optimisation for real-time deployment on edge devices.