Overview
YOLO-based detection paired with a classifier workflow to distinguish valid targets from non-targets and background in aerial imagery/video. This page is meant to hold detailed methods, dataset notes, and demo results.
Key highlights
A quick snapshot of the most important features and design goals.
- Real-time detection pipeline (segmentation/detection front-end).
- Coarse “valid/invalid” classification stage for decision-making.
- Designed for practical deployment with OpenCV / video streams.
Links
Primary references and artifacts for this project.