Vehicle speed dataset for the major European road network derived from Sentinel-2 imagery, 2022-2026
This paper introduces a continental-scale dataset of individual vehicle speeds on major European E-roads for 2022–2026, derived by analyzing inter-band displacements of moving vehicles in Sentinel-2 satellite imagery to support traffic pattern analysis and transport modeling.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
To understand how people move across a continent, we usually need to count them. For decades, traffic engineers have relied on sensors buried in the asphalt or cameras mounted on poles to measure how fast cars are traveling. These tools are precise, but they only see the road directly beneath them. If a car is driving on a remote highway in Norway or a secondary road in Poland, it might go entirely unnoticed. Other methods, like using data from navigation apps on drivers' phones, offer a wider view, but they only capture the vehicles of people who have agreed to share their location, often skewing the picture toward specific groups or regions. For a complete, unbiased map of how traffic actually flows across Europe, scientists have long needed a way to watch the roads from above without needing permission from the drivers or installing new hardware on the ground.
A team of researchers has now provided exactly that, creating a massive new dataset that tracks vehicle speeds across the major European road network using nothing but satellite imagery. By analyzing images taken by the European Space Agency's Sentinel-2 satellite, they have mapped the speed of individual vehicles on motorways, trunk roads, and primary and secondary routes from 2022 through 2026. This work transforms a standard photograph into a motion picture, allowing scientists to see not just where cars are, but how fast they are moving, all without ever touching the road.
The secret to this discovery lies in a quirk of how modern satellites take pictures. Most satellites capture all the colors of light—blue, green, and red—at the exact same moment. However, the Sentinel-2 satellite uses a different method called a push-broom scanner. Instead of snapping a single instant, it scans the ground line by line, capturing the blue light, then the green, and finally the red, in rapid succession. Because the satellite is moving at thousands of kilometers per hour, there is a tiny fraction of a second between when the blue sensor sees a specific spot on the ground and when the red sensor sees it. If a car is driving fast enough during that split second, it will appear in slightly different positions in the blue, green, and red layers of the same image.
The researchers built a sophisticated computer program to hunt for these tiny shifts. When a vehicle moves, it leaves a faint "echo" across the three color bands, appearing as a bright spot that has drifted just a few pixels from one color to the next. The software scans millions of road segments, looking for these displaced peaks of light. Once it finds a match, it links the three positions together to form a short path. By knowing exactly how much time passed between the blue and red images, and measuring how far the car moved in that time, the system calculates the vehicle's ground speed. This process turns a static image into a dynamic measurement, revealing the speed of individual cars as they travel along the road.
The resulting dataset is enormous, containing nearly 22 million records of individual vehicle observations across Europe. Each record includes the path the car took, its estimated speed, the direction it was heading, and the exact time the satellite took the picture. The data covers the years 2022 to 2026 and focuses on the most important roads in the European network, including motorways and major highways, as defined by the open-source map database OpenStreetMap. The researchers verified their method by comparing the satellite measurements against known physical limits and running simulations to ensure the numbers make sense. They found that while the method is incredibly powerful, it has natural limits: it works best for cars moving at moderate to high speeds and can struggle to detect very slow vehicles or those that blend in too well with the road surface.
This work represents a significant shift in how we observe traffic. Unlike traditional methods that depend on fixed sensors or voluntary data sharing, this approach relies entirely on the satellite's own movement and the physics of light. It offers a consistent, continent-wide view that is not biased by where sensors happen to be installed or which drivers choose to share their data. The researchers have made the entire dataset and the code used to create it freely available to the public. This means that urban planners, climate scientists, and emergency responders can now study traffic patterns, model emissions, or plan evacuation routes using a source of information that is independent, reproducible, and covers the entire European road network. The data captures a specific slice of time—mostly late morning hours when the satellite passes overhead—but it provides a foundational layer of truth about how vehicles move across the continent that was previously impossible to gather at this scale.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.