Exploring Techniques for Poverty Level Prediction from Satellite Imagery

Determining the poverty levels of various regions throughout the world is crucial in identifying interventions for poverty reduction initiatives and directing resources fairly. However, reliable data on global economic livelihoods is hard to come by, especially for areas in the developing world, hampering efforts to both deploy services and monitor/evaluate progress. Satellite images can be used to detect economic activity and infrastructure of a region in order to predict the wealth index (or poverty level) of a location. Research on poverty prediction from satellite images using deep learning has shown promising results. By leveraging the rich spatial information captured by satellite imagery, deep learning models can analyze various features and patterns to predict poverty levels. The aim of this exploratory survey is to comprehend and evaluate the work done in the field of poverty prediction and to understand how data captured using satellite can be mapped to various wealth and socio-economic wellbeing parameters. In this survey we examine different satellite image datasets that can be utilized for predicting poverty levels, while also reviewing previous research conducted in this domain. Ultimately, we propose an architecture that shows promise in delivering favourable outcomes for this task.