Using high-resolution remote sensing data to study slum growth and population clustering, Delhi

Abstract

This study investigates slum growth and population clustering in Delhi using high-resolution remote sensing and advanced spatial analysis techniques. Very-high-resolution satellite imagery, combined with convolutional neural network segmentation and multi-temporal change detection, enabled detailed mapping of informal settlement morphology and identification of densification, expansion and consolidation patterns. Population distribution was estimated through structural proxies derived from building footprints and density metrics, revealing significant intra-settlement variation and highlighting areas of heightened vulnerability. The methodological framework demonstrates the value of integrating machine learning, ancillary spatial datasets and uncertainty assessment to produce reliable, policy-relevant insights. The findings underscore the importance of fine-scale geospatial intelligence for equitable urban planning, particularly in rapidly evolving megacities where conventional data sources are limited or outdated.