NASA Harvest

Machine Learning

Scalable Cropland Mapping Workshop

Monday August 22, 2022 - Thursday August 25, 2022
In person: Room 301 at 4600 River Road, Riverdale Park, MD
Virtual: join through https://umd.zoom.us/my/izvonkov
Public Google Doc: link
Public Google Drive: link

This workshop will do a deep dive into scalable cropland mapping with deep learning. Fundamental topics in machine learning, satellite time series data, labeling, and scalable systems will be covered. Every participant will generate their own project for cropland mapping with deep learning.

Schedule

All times are in local time for Maryland.

	Monday August 22, Day 1/4 - Scalable Cropland Mapping
Time	Activity	Instructor
13:00 - 13:40	Intro to workshop	Ivan
13:40 - 15:00	OpenMapFlow preview	Ivan
15:00 - 15:30	Tea break
15:30 - 17:00	Labeling in CollectEarthOnline	Ivan

	Tuesday August 23, Day 2/4 - Scalable Cropland Mapping
Time	Activity	Instructor
9:00 - 9:30	Labeling in CollectEarthOnline	Ivan
9:30 - 10:00	ML models/intro	Gabi
10:00 - 10:30	Labeling in CollectEarthOnline	Ivan
10:30 - 11:00	Tea break
11:00 - 11:30	Label Consensus	Hannah
11:30 - 12:00	ML for Earth Observation data	Hannah
12:30 - 13:30	ML for Earth Observation data discussion	Ivan, Gabi
13:30 - 15:00	Lunch
15:00 - 15:30	Machine learning systems	Ivan
15:30 - 17:00	Generating a new OpenMapFlow project	Ivan

	Wednesday August 23, Day 3/4 - Scalable Cropland Mapping
Time	Activity	Instructor
09:00 - 09:30	Generating ML ready data	Ivan
09:30 - 10:30	ML models for EO data	Gabi
10:30 - 11:00	Tea break
11:00 - 12:30	Generating ML ready data revisited	Ivan
12:30 - 14:00	Lunch
14:00 - 15:00	Training a model	Ivan
15:00 - 16:00	Making a map	Ivan
16:00 - 17:00	Discussion / Tea break (outside)

	Thursday August 24, Day 4/4 - Scalable Cropland Mapping
Time	Activity	Instructor
9:00 - 10:00	Improving Map Discussion	Ivan
10:00 - 10:30	Improving Map In Practice	Ivan
10:30 - 11:00	CropHarvest	Gabi
11:00 - 11:30	Tea break
11:30 - 12:00	Harvest's crop type mapping with TIML	Gabi
12:00 - 13:00	Conclusion and discussion	Ivan

Instructors
Ivan Zvonkov (University of Maryland)	Gabriel Tseng (McGill University)	Dr. Hannah Kerner (Arizona State University)	Dr. Catherine Nakalembe (University of Maryland)

Additional Resources

• NASA Harvest Github organization: https://github.com/nasaharvest
• OpenMapFlow Github repository: https://github.com/nasaharvest/openmapflow
• Example W&B Project: https://wandb.ai/nasa-harvest/crop-mask-example
• CropHarvest Github repository https://github.com/nasaharvest/cropharves
• Introduction to Earth Observation (curated list of useful resources)
• NASA Harvest website: https://nasaharvest.org
• Radiant Earth ML Hub: https://mlhub.earth/
• Google Earth Engine: https://earthengine.google.com/
• Rußwurm M., Körner M., Self-attention for raw optical Satellite Time Series Classification (2019), link
• Stehman, S. V., & Foody, G. M. (2019). Key issues in rigorous accuracy assessment of land cover products. Remote Sensing of Environment, 231, 111199.
• Kerner, H. R., Tseng, G., Becker-Reshef, I., Barker, B., Munshell, B., Paliyam, M., and Hosseini, M. (2020). Rapid Response Crop Maps in Data Sparse Regions. In Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining Workshops, link.
• Rustowicz, R., Cheong, R., Wang, L., Ermon, S., Burke, M., and Lobell, D. (2019). Semantic segmentation of crop type in africa: A novel dataset and analysis of deep learning methods. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (pp. 75-82), link.
• Tseng, G., Kerner, H., and Rolnick, D. (2022). TIML: Task-Informed Meta-Learning for Agriculture. arXiv preprint arXiv:2202.02124, link.
• Wang, S., Waldner, F., and Lobell, D. B. (2022). Unlocking large-scale crop field delineation in smallholder farming systems with transfer learning and weak supervision. arXiv preprint arXiv:2201.04771, link.
• You, J., Li, X., Low, M., Lobell, D., and Ermon, S. (2017). Deep Gaussian Process for crop yield prediction based on remote sensing data. In Thirty-First AAAI Conference on Artificial Intelligence, link.
• Kerner, H. R., Rebbapragada, U., Wagstaff, K. L., Lu, S., Dubayah, B., Huff, E., Raman, V., and Kulshrestha, S. (2022). Domain-Agnostic Outlier Ranking Algorithms—A Configurable Pipeline for Facilitating Outlier Detection in Scientific Datasets. Frontiers in Astronomy and Space Sciences, 9, 867947, link.
• Tusubira, J. F., Akera, B., Nsumba, S., Nakatumba-Nabende, J., and Mwebaze, E. (2020). Scoring root necrosis in cassava using semantic segmentation. In CVPR Workshops 2020, link.

Supported by