Kreike Awarded Microsoft AI for Earth Grant

Posted
April 23, 2021
Present day satellite image of 1943 area

Emmanual Kreike's project, “Proof of Concept Project to Extract Land Use Data from a Historical 1943 Orthomosaic through Machine Learning,“ has been selected for an AI for Earth Microsoft Azure Compute Grant. The award grants credits for Azure consumption up to $10000 USD, to be utilized within one year.

The objective is to assess the potential of machine learning assisted extraction of environmental and land use data from time series of aerial photos from 1943-1972. This serves as a first phase for a larger project that aims to develop a machine learning assisted methodology and tool set to extract environmental data from large time series of aerial photographs. At present, older series of aerial photos are rarely used and only in very small quantities because they must be analyzed manually. If successful, with the new algorithm, the time depth of analyzing land use and environmental change using bird’s eye view images can be pushed back from the 1980s (when detailed SPOT satellite images become available) to the 1930s (when detailed aerial photography becomes widely available) allowing for the identification of longer-term global environmental trends and climate change.

Kreike is collaborating on this project with Wangyal Tsering Shawa and William Guthe. They will be hiring several students this summer to work with them on the project during the summer and fall.

Emmanuel Kreike is Professor of History at Princeton University. He has conducted extensive research on environmental change and land use in southern Africa, in particular in the Namibian-Angolan border region captured in the aerial photo time series that the project is using. He wrote three books on the environmental history of the study area and can visually interpret the patterns and objects on the time series of images. He has worked with manual and visual analysis of aerial photos and historical maps for over 20 years.

William Guthe is the Senior Geographical Information Systems (GIS) Visualization Analyst in Research Computing at the Office of Information Technology, Princeton University. With Tsering Shawa he has taught “GIS for Public Policy” to graduate students in the Princeton School of Public and International Affairs for ten years and has co-taught an undergraduate Digital and Spatial History Lab four times with Tsering Shawa and Emmanuel Kreike.

Tsering Wangyal Shawa is the GIS and Map Librarian and Head, Map and Geospatial Information Center at the Peter B. Lewis Science Library at Princeton University and has been researching and teaching GIS through workshops and classes for 20 years, including “GIS for Public Policy” and HIS 278 Digital and Spatial History.


Image credit: Esri Imagery Hybrid basemap