Taking the Long View: Enhancing Learning On Multi-Temporal, High-Resolution, and Disparate Remote Sensing Data for Measuring the Spread of Buildings
Abstract:
The progress made in computer vision and satellite technology has opened up new possibilities for observing societies and infrastructure. Powered by a lengthening historical record of high-resolution satellite imagery and an ever-growing set of labels for training machine learning models, it is becoming feasible to investigate a decade of changes in built infrastructure and landscapes. These analyses can assist decision-makers with valuable insights into population shifts, economic trends, and infrastructure performance. Nevertheless, challenges inherent with this kind of imagery — such as varying image quality, imbalances in data collection between urban and rural areas, high costs, and the absence of image metadata — can impede the use and efficacy of these methods.
In this talk, I will introduce our team’s work in developing techniques to measure the changes in built infrastructure in developing regions of sub-Saharan Africa. Specifically, I will discuss challenges we faced in working with high-resolution multi-temporal satellite imagery towards the task of detecting the emergence of buildings in rural Kenya over time. I will then present results on a case study analyzing the relationship between the expansion of electricity access and the growth in human settlements over a decade. Our non-intuitive and statistically robust finding challenges conventional wisdom about infrastructure provision and rural-urban migration, with potentially broad implications for assessing the impacts of infrastructure investments on rural lives and livelihoods. I will also briefly outline related work from our team on multi-temporal imagery that considers migration to areas with climate hazards and longitudinal measurement of road quality.
Jay Taneja is an Assistant Professor in the Manning College of Information and Computer Sciences at the University of Massachusetts, Amherst. As the lead of the STIMA Lab, he and his team develop and study applications of sensing and communications technology on the measurement and planning of societal-scale infrastructure systems in developing regions. He also leads the e-GUIDE Initiative, a multi-university consortium that collaborates with partners across sub-Saharan Africa with an aim to transform the approaches used for planning and operations of electricity infrastructure in the region. Prior to joining UMass, he was a Research Scientist leading the Energy team at the IBM Research - Africa lab in Nairobi, Kenya, from 2013 to 2016.
