Agriculture
This page is about the intersection of agriculture and machine learning in the context of climate change mitigation. For an overview of agriculture as a whole, please see the Wikipedia page on this topic.
As described in the paper "Tackling Climate Change with Machine Learning"[1]:
Agriculture is responsible for about 14% of GHG emissions[2]. This might come as a surprise, since plants take up CO2 from the air. However, modern industrial agriculture involves more than just growing plants. First, the land is stripped of trees, releasing carbon sequestered there. Second, the process of tilling exposes topsoil to the air, thereby releasing carbon that had been bound in soil aggregates and disrupting organisms in the soil that contribute to sequestration. Finally, because such farming practices strip soil of nutrients, nitrogen-based fertilizers must be added back to the system. Synthesizing these fertilizers consumes massive amounts of energy, about 2% of global energy consumption[3]. Moreover, while some of this nitrogen is absorbed by plants or retained in the soil, some is converted to nitrous oxide, a greenhouse gas that is about 300 times more potent than CO2.
Such industrial agriculture approaches are ultimately based on making farmland more uniform and predictable. This allows it to be managed at scale using basic automation tools like tractors, but can be both more destructive and less productive than approaches that work with the natural heterogeneity of land and crops. Increasingly, there is demand for sophisticated tools which would allow farmers to work at scale, but adapt to what the land needs. This approach is often known as “precision agriculture.”
Machine Learning Application Areas
- Precision agriculture: Typical industrial agriculture releases CO2 into the atmosphere by disrupting natural soil chemistry and biodiversity, and also requires chemicals that are emissions-intensive both to produce and to use. ML can help monitor emissions; reduce the need for chemicals by pinpointing pests, diseases, and weeds; and change agricultural paradigms by controlling physical robots.
- Food security: By affecting rainfall and the timing of growing seasons, climate change poses a risk to food security. Machine learning can support information gathering around food supply chains, providing early warnings about -- and triggering preventative action around -- famines.
- Monitoring agricultural emissions: Agriculture is a major contributor to greenhouse gas emissions via methane (in particular from cattle farming) and nitrous oxide (from fertilizer), in addition to carbon dioxide from soil carbon breakdown. Machine learning can help track agricultural emissions directly, in addition to monitoring crop cover and livestock to help assess emissions potential.
Background Readings
- Characterizing agricultural drought in the Karamoja subregion of Uganda with meteorological and satellite-based indices
- Remote Sensing Techniques for Soil Organic Carbon Estimation: A Review
Online Courses and Course Materials
Conferences, Journals, and Professional Organizations
Journals
- Computers and Electronics in Agriculture: International journal covering computer hardware and software for solving problems in agriculture, agronomy and horticulture. Website here.
- Precision Agriculture: International Journal on Advances in Precision Agriculture. Website here.
Workshops
- ICLR 2020 Workshop on Computer Vision for Agriculture: Exposes the progress and unsolved problems of computational agriculture to the AI research community. Website here.
- CVPR 2020 Workshop and Prize Challenge on Agriculture-Vision: Present recent progress on computer vision research for tackling impactful challenges in agriculture. Website here.
Groups and Labs
- Australian Centre for Field Robotics: Robotic institute focusing on autonomous robots that can work in outdoor environment. Website here.
Library and Tools
Some packages for working with remote sensing data are,
- eo-learn: A python package maintained by the European Space Agency, giving easy access to imagery from Sentinel satellites, as well as utilities for data processing. Github, medium-post.
Data
Remote sensing
Remote Sensing Datasets offers great opportunity to monitor agriculture and can be georeferenced to match ground measurements. In particular, Google earth engine offers a convenient interface over freely available satellite imagery such as Landsat and Sentinel.
Remote Crop Identification
- Agriculture-Vision: A Large Aerial Image Database for Agricultural Pattern Analysis: A dataset composed of 94,986 aerial images from 3,432 farmlands across the United States. The images contains RGB channels and Near-infrared at a resolution of 10 cm per pixel. Dataset, and paper describing the dataset.
Kaggle datasets
- Plant Seedlings Dataset
- Agriculture crop images
- Global Food & Agriculture Statistics
- Weed Detection in Soybean Crops
- crop and weed detection data with bounding boxes
- DeepWeedsX
- California Crop Mapping - 2014
References
- ↑ 1.0 1.1 Rolnick, David; Donti, Priya L.; Kaack, Lynn H.; Kochanski, Kelly; Lacoste, Alexandre; Sankaran, Kris; Ross, Andrew Slavin; Milojevic-Dupont, Nikola; Jaques, Natasha; Waldman-Brown, Anna; Luccioni, Alexandra (2019-11-05). "Tackling Climate Change with Machine Learning". arXiv:1906.05433 [cs, stat].
- ↑ Intergovernmental Panel on Climate Change (2014). Climate Change 2014 Mitigation of Climate Change: Working Group III Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press. doi:10.1017/cbo9781107415416. ISBN 978-1-107-41541-6.
- ↑ Montoya, Joseph H.; Tsai, Charlie; Vojvodic, Aleksandra; Nørskov, Jens K. (2015-06-10). "The Challenge of Electrochemical Ammonia Synthesis: A New Perspective on the Role of Nitrogen Scaling Relations". ChemSusChem. 8 (13): 2180–2186. doi:10.1002/cssc.201500322. ISSN 1864-5631.