How many times have someone sent you their live location on Maps, and you’ve ended up playing “Where’s Waldo?”, and have a hard time finding the place? A new 3D map that exclusively maps buildings can change that. GlobalBuildingAtlas has generated 3D models of 97% of all buildings, using satellite imagery and machine learning. Around 800,000 satellite images from 2019 were used to map the area, volume, and height of the buildings.
The dataset, published in the open-access journal Earth System Science Data on 1 December, covers 2.75 billion buildings, including heights and footprints at a spatial resolution of 3 metres by 3 metres.
According to Xiaoxiang Zhu, an Earth observation data scientist at the Technical University of Munich in Germany, the 3D map can support climate modelling, urban planning, and disaster risk assessment. United Nations (UN) Sustainable Development Goals in cities and communities can benefit from this model, he adds.
It is challenging to create a 3D map of buildings because it needs state-of-the-art laser scanning techniques or high-resolution stereo imagery. This model uses laser scanning and deep learning techniques. To train the model, LiDAR (light detection and ranging) reference data from 168 cities, primarily in Europe, North America, and Oceania, were used.
Asia tops the chart with 1.22 billion structures with 1.27 trillion cubic meters of building volume, a proof of steep urbanization and dense populations in countries such as China, India, and Southeast Asia. Africa ranks second with 540 million buildings; however, the combined volume is around 117 billion cubic meters, indicating the prevalence of small, low-rise structures.
City-scale analyses indicate the correlation between building space and population density and economic development. Finland has six times the building volume of Greece, while Niger’s is 27 times the global average. This provides better insight into measures of urban growth, which traditional 2D resources can obscure. Disaster management research has so far relied on 2D data, which lacks this level of precision; a 3D model could make datasets more accurate to work with.
Dorina Pojani, an urban planning researcher at the University of Queensland in Brisbane, Australia, says, “Since this can be regularly updated, it will be very valuable over the next five to ten years, as the data set will reveal how urban areas develop over time.”
The datasets can also uncover corruption in how land is used by people in high places, and whether they have their houses strategically built for personal gain. In fact, previous research has linked informal settlements to election outcomes, with political parties often paying no heed to these settlements.
Who thought that unfolding an urban map can become similar to a pop-up card? Now, we have it.
Copyright @smorescience. All rights reserved. Do not copy, cite, publish, or distribute this content without permission.
