Using EOLearn to build a machine learning pipeline
summary
Claim: a reproducible pipeline using EOLearn and Sentinel-2 multispectral data can detect floating macroplastics at sub‑pixel scale by combining the floating debris index and NDVI, a local normalization moving‑window correction, and scikit‑learn classification. The team builds EO tasks, a training and prediction workflow, and a deployable detector that scales via Sentinel Hub. Results show about 88% precision with low recall; code and tasks are open sourced and improve generalization across regions.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:15Satellite data for ML
- 8:49Plastic detection from space
- 15:47Scaling challenges
- 17:12EOLearn for Earth observation
- 23:52Training and prediction pipelines
- 25:36Fixing overfitting
- 30:49Results and future steps
- 32:14Q&A
related talks
19mDigital Earth Pacific DEP cloud native approaches to detect mineral extraction
36mClassification of Satellite Imagery
31mFrom Radar Echoes To Real-Time Recognition: Deep Learning on Another Planet -An Open Source Workflow
30mAutomated river plastic monitoring Using..
45mFrom Radar Echoes To Real-Time Recognition: Deep Learning on Another Planet an Open Source Workflow
46mML Use Cases at Think Big Analytics
25mData science and machine learning in wastewater intelligence 🌳
18mData science and machine learning in wastewater intelligence