• P Bharath
• K R Sumana

This project aims to investigate the use of deep learning techniques, specifically Recurrent Neural Networks (RNNs), to predict the poverty level of a region using satellite imagery. The hypothesis is that by leveraging the information captured in daytime and nighttime satellite images, it is possible to reliably estimate the wealth index of a city. To conduct this research, a dataset of 88,386 images from 44,193 cities in Africa, South America, Asia, Europe, and the Caribbean was compiled. Each city has both daytime and nighttime satellite images and a corresponding wealth index. RNNs will be trained to learn the relationship between these images and the wealth index, capturing indicators of poverty and wealth. The models are evaluated for predictive accuracy. This study aims to demonstrate the effectiveness of deep learning and satellite imagery in poverty prediction, with potential benefits for policymakers, NGOs, and researchers. The approach could aid in developing targeted interventions and resource allocation in areas lacking ground-level data.

deep learning, poverty prediction, satellite imagery, Recurrent Neural Networks (RNNs), wealth index.

"Geospatial poverty forecasting with satellite data", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 7, page no.h122-h126, July-2024, Available :http://www.jetir.org/papers/JETIR2407716.pdf

"Geospatial poverty forecasting with satellite data", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 7, page no. pph122-h126, July-2024, Available at : http://www.jetir.org/papers/JETIR2407716.pdf