Abstract
Computer vision for large scale building detection can be very challenging in many environments and settings even with recent advances in deep learning technologies. Even more challenging is modeling to detect the presence of specific buildings (in this case schools) in satellite imagery at a global scale. However, despite the variation in school building structures from rural to urban areas and from country to country, many school buildings have identifiable overhead signatures that make them possible to be detected from high-resolution imagery with modern deep learning techniques. Our hypothesis is that a Deep Convolutional Neural Network (CNN) could be trained for successful mapping of school locations at a regional or global scale from high-resolution satellite imagery. One of the key objectives of this work is to explore the possibility of having a scalable model that can be used to map schools across the globe. In this work, we developed AI-assisted rapid school location mapping models in eight countries in Asia, Africa, and South America. The results show that regional models outperform country-specific models and the global model. This indicates that the regional model took the advantage of having been exposed to diverse school location structure and features and generalized better, however, the global model was the worst performer due to the difficulty of generalizing the significant variability of school location features across different countries from different regions.
Highlights
Reliable and accurate data about school locations have become vital to many humanitarian agencies and governments to effectively plan, manage, and monitor the provision of quality education and learning in accordance to the UN sustainable development goal 4(SDG4 [1]) that ensure equal access to opportunity (SDG10 [1]) [2]
The East African regional model was trained with Kenya and Rwanda datasets, and the West African regional model was trained with Niger, Sudan, Mali, Chad, and Sierra Leone datasets
The regional models outperformed the global and the country-specific models, which indicates that the models were exposed to more diverse school features affirming the fact that the greater variability in the dataset the better the models
Summary
Reliable and accurate data about school locations have become vital to many humanitarian agencies and governments to effectively plan, manage, and monitor the provision of quality education and learning in accordance to the UN sustainable development goal 4(SDG4 [1]) that ensure equal access to opportunity (SDG10 [1]) [2]. Understanding the location of schools can help governments and international organizations gain critical insights into the needs of vulnerable populations, and better prepare and respond to exogenous shocks such as disease outbreaks or natural disaster development programs aiming to provide internet connection to schools in developing countries which requires accurate and comprehensive datasets of school locations [2]. Visual perception to extract feature hierarchies and generalization ability is enhanced on several levels [12]. These algorithms have shown that traditional techniques are slow and erroneous; they require extensive post-processing to differentiate infrastructure [17]. DNN classifiers that work beyond task-based methods for object recognition and can carry out adaptive and deep learning from multi-resolution imagery for object detection
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
