Abstract
During the last 20 years, fast urbanization activities have been highly concentrated in just few countries (e.g., China, India, and Nigeria) and have led to the emergence of large urban aggregations, with high population density. Still, very few researches have focused on this dynamic phenomenon with a global perspective using multisource remote sensing data. In this article, combining radar and spectral sensors of different spatial resolution, a novel approach based on a novel hierarchical biclustering technique is proposed and proved to be effective in discriminating the underlying change patterns without pre-estimating the number of clusters. To this aim, experimental results focused on newly emerging megalopolis in China, India, and Nigeria, as well as on the highly urbanized and stable Lombardy region in Italy, are presented. The analysis of the results allows us to understand, in a global and comparative perspective, the spatiotemporal differentiation of urban density and how cities are changing and evolving in the building volume and, to some extent, their economic level.
Highlights
In the last few decades, the emergence of megalopolises or megacities has already attracted great attention in investigating urbanization activities and planetary-scale changes issues [1, 2]
According to the ”World Urbanization Prospects 2018” report from the United Nations, projections show that another 2.5 billion people will migrate from rural to urban areas by 2050 and nearly close to 90% of this expected increment is highly concentrated in just a few countries (India, China and Nigeria) rather than in the most urbanized regions [7]
We focus on two megalopolises in China, namely Jingjinji (JJJ) and Yangtze River Delta (YRD), one megalopolis in India, namely the National Capital Region
Summary
In the last few decades, the emergence of megalopolises or megacities has already attracted great attention in investigating urbanization activities and planetary-scale changes issues [1, 2]. Nighttime lights are a proxy to the economic development, the population density, and quite a few more social parameters [18, 19] By using both data sets we expect to be able to jointly consider changes due to urban expansion and those due to increasing socio-economic activities. Following [21], a “megalopolis” is a cluster of multiple urban areas where usually the government policy aims at knitting the area together and promoting development through transportation and communication links Since these areas are very large, and temporal patterns are a a priori unknown, unsupervised techniques are needed to analyze remote sensing data at their scale. In our previous work [20], the joint use of heterogeneous sensors allows an unsupervised discover of spatio-temporal features and deeper relationships between urban constructions and nighttime-light changes, which reveal the connections between built-up area changes and socio-economic development.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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
