Novel deep learning and big data analytics can play a crucial role unveiling detailed information about urban structures from space.
Understanding urbanization is key to answering some of the major challenges that society faces today, and these include major themes such as climate change and global inequalities. Overcoming these challenges lie at the heart of international endeavors such as the United Nations’ call for “Sustainable Cities and Communities”. But can deep learning tools be used to better understand the complex phenomenon of urbanization? The answer is yes: a research team led by Xiaoxiang Zhu, Helmholtz AI steering board member and head of department at the German Aerospace Center’s Remote Sensing Technology Institute, showed that novel deep learning and big data analytics can play a crucial role unveiling detailed information about urban structures from space.
In the paper published in Remote Sensing of Environment, the team has made available the first-ever global and quality controlled urban local climate zones classification derived from satellite images covering all cities across the globe with a population greater than 300,000. This is especially important because such data was not available for many developing countries that are particularly vulnerable to issues of climate change and global inequalities. “We hope the open access of our dataset will encourage research on a better understanding of the global change process of urbanization,” says Zhu. “The aim is that researchers from across disciplines will use this data in their work. And that the dataset can serve stakeholders such as the United Nations to improve their spatial assessments of urbanization,” she adds.
The paper can be accessed at: https://www.sciencedirect.com/science/article/pii/S0034425721005149