The United Nations hold well-developed and safe roads as an important part of infrastructure in their Sustainable Development Goals. Until now, however, there has been no benchmark for the condition and state of road networks worldwide. Researchers at the Institute of Geography of Heidelberg University and HeiGIT (Heidelberg Institute of Geoinformation Technology) have addressed this gap by using artificial intelligence and satellite imagery to create an open-access dataset that maps and classifies more than nine million kilometers of roads worldwide. The dataset also captures changes in road conditions over time, providing valuable information for humanitarian applications and serving as an important indicator for assessing socioeconomic development, particularly in data-scarce regions.
The dataset is based on high-resolution images of the Earth's surface captured by the PlanetScope satellites between 2020 and 2024. Using deep learning, researchers in Heidelberg led by Prof. Dr. Alexander Zipf were able to identify approximately 9.2 million kilometers of major thoroughfares and classify them based on their pavedness and width. "Our model is about 20 percentage points more accurate than previously used datasets and makes it possible to track changes in road infrastructure over periods of several years," explains Dr. Sukanya Randhawa, who leads the "GeoAI for Good" projects at HeiGIT. In particular, this makes it possible to draw conclusions about how passable a road is under changing conditions – such as during extreme weather events. To this end, the researchers have converted the data into a Humanitarian Passability Matrix, which can be used to plan humanitarian operations.
Road pavedness can also provide insight into the level of development of individual countries or regions. As the analyses for 2020 through 2024 show, the degree of pavedness is directly linked to level of development and development potential. For instance, countries with a relatively high proportion of paved roads rank higher on the United Nations' Human Development Index than countries where major roadways are unpaved and road-building projects are progressing more slowly. "Road infrastructure can therefore serve as an indicator of socioeconomic progress at both the global and local levels, particularly in regions where traditional development indicators, such as nighttime light data, reach their limits," emphasizes Sukanya Randhawa. This is demonstrated by a case study from Ghana, among others. The data reveals infrastructure gaps in specific neighborhoods of the capital, Accra, and can thus contribute to more equitable urban planning and resource allocation.
The dataset is publicly available via a platform operated by the United Nations Office for the Coordination of Humanitarian Affairs. According to Prof. Zipf, the data is continuously updated. "They show that the global road network can serve as a dynamic measure of change and thus as a basis for various areas of action – ranging from scientific applications and the planning of humanitarian missions to concrete investment decisions for economic development," says the researcher, who heads the Geoinformatics department at the Institute of Geography of Heidelberg University and HeiGIT, which is supported by the Klaus Tschira Foundation.
The results of the research were published in the journal Nature Communications.
