There are many unobservable or unrecordable variables in the agriculture-climate-human cycle, which poses big challenges for scientists to estimate impacts and predict the future trends in crop production, food supply, market pricing, etc. Artificial intelligence gives us hope to skip those variables and sophisticated simulation equations by building direct relationship models between the variables which we do have observations and the information which could be used to advise farming activities, predict food market, or making related policies. We believe that once these AI models are built and trained using historical datasets, agricultural stakeholders would be able to get more direct, quick, high-quality information to make more accurate assertion about the history, present and future development of agricultural geography. Meanwhile, we also admit that AI models are not in perfection and it takes a lot of engineering to get them trained to meet the qualification for operationally serving in real-world scenarios. Therefore, we call for the community efforts to discuss about the theory, methodology and techniques to achieve this goal together.
This session calls for work related to artificial intelligence or agricultural geography, including but not limited to:
Agriculture-climate remote sensing
Agriculture-climate social science
AI product provenance and workflow
AI tooling and services for agriculture-climate datasets
Geospatial Cyberinfrastructure for AI (CI for AI)
If interested in talking in this session, please send your abstract and the Personal Identification Number (PIN) for AAG 2020 to Ziheng Sun (George Mason University; email@example.com) and Manzhu Yu (Penn State University; firstname.lastname@example.org).
Human geography is historically at some level dominated by agricultural environment and climate factors. Even today the agricultural environment and climate still have a wider and more direct effect on the human society. Farming on different types of soils in various climates requires different set of knowledge and skills to secure the yield and quality. Also, human factors, policy, costs, and prices cannot be ignored in explaining the farming activities in today’s world agriculture landscape. The environmental and human factors interfere with each other and results in the current crop distribution, food chain, agricultural market distribution and pricing. To get a better understanding of agricultural geography, we must build models/tools/systems to help estimate the impacts of the changes of each component in the agriculture-climate-human networks. However, the mutual interfering is very difficult to simulate by traditional numeric modeling methods due to the lack of fine-resolution observations and the complexity of the underlying mechanisms. Artificial intelligence, however, lets us see some possibility to simulate these relationships without diving into the ultra-complicated numeric models. We have seen several successful application of using the state-of-art AI techniques in accomplishing difficult tasks in agriculture, e.g., automatically crop mapping at large scale, which can only imagine before. We would like to invite community geographers to attend us to talk about your opinions/experiences/visions on using artificial intelligence in analyzing/monitoring/predicting the relationships and consequences in agriculture and climate-related geographical factors.
|Presenter||Bing Lu*, University of Toronto - Mississauga, Yuhong He, University of Toronto Mississauga, Jiali Shang, Agriculture and Agri-Food Canada, Jiangui Liu, Agriculture and Agri-Food Canada, Cameron Proctor, University of Windsor, Phuong Dao, University of Toronto Mississauga, Ali Al Wafi, University of Toronto Mississauga, Monitoring Crop Physiological Status Using Airborne Hyperspectral and Thermal Images||15|
To access contact information login