Classification of Minor Roads Inside Dense Informal Settlements of Mumbai Using Machine Learning

The paper discusses the issue of haphazard development in Indian cities, leading to the rise of unplanned informal settlements and narrow roads, which makes it difficult for essential services to be delivered to residents. The lack of accurate network information on minor roads inside informal settlements further exacerbates this problem. The paper proposes the use of a deep neural network (DNN) model to classify urban features, including minor roads inside dense informal settlements of Mumbai city, using high-resolution satellite imagery. The proposed model achieved an accuracy of around 95% in classifying urban features and 85% in classifying minor roads inside informal settlements. The outcomes presented in the paper have significant implications for sustainable development goals, as they can support effective resource planning to cater to marginalized communities living in slums. Accurately mapping minor roads and urban features can help improve the delivery of essential services such as emergency support, healthcare, and sanitation to these communities. The proposed approach can be replicated in other cities facing similar challenges, contributing to marginalized communities’ overall development and well-being.