Dutch Researchers Strap Thermal Cameras to Bikes, Building Spectrascapes to Map Urban Heat
A bike-mounted rig captured 17,718 calibrated RGB, near-infrared, and thermal images across Dutch towns and cities, filling a gap satellites and Google Street View cannot.
Takeaways
- Spectrascapes releases 17,718 calibrated street-level images with RGB, near-infrared, and thermal channels captured across four Dutch urban areas
- Sensors are mounted on bicycles, filling the gap between satellite overhead views and standard Google Street View RGB imagery
- Claimed to be the first open-access multi-spectral terrestrial-view dataset, with full hardware and calibration methodology published
- Demonstrated use case reconstructs 3D street geometry from only three RGB frames using bike GPS displacement as baseline
- Enables street-level urban heat island analysis, tree-shadow valuation, and fusion with LiDAR maps like the Netherlands AHN
- Limited to Dutch geography and modest scale, but the open hardware recipe invites community replication in other cities
Satellite imagery gives you cities from above. Google Street View gives you sidewalks in plain RGB. Neither tells you how hot a specific brick facade runs at 3pm, or how much chlorophyll a curbside tree carries. A new open dataset from researchers in the Netherlands tries to close that gap by strapping thermal and near-infrared cameras to bicycles and riding them through four Dutch urban areas.
Spectrascapes contains 17,718 street-level multi-spectral images captured with RGB, near-infrared, and thermal sensors mounted on bikes, spanning a village, a town, a small city, and a larger urban area. The authors claim it is the first open-access dataset of its kind, with heavy emphasis on calibration and data quality.
RGB leaves too much on the table
Cities are getting hotter and denser, and the existing monitoring tools all have blind spots. Manual inspections don't scale. Embedded sensors give inconsistent spatio-temporal coverage. Remote sensing is stuck overhead. Standard street-view imagery captures only the visible spectrum.
Satellites see rooftops but miss shaded courtyards. Fixed weather stations produce sparse point measurements. Google Street View cars discard the thermal and near-infrared signals that carry the most useful information about vegetation health, surface temperature, and building materials. Near-infrared reflectance feeds vegetation indices like NDVI, and thermal imagery is how you actually measure urban heat island effects at ground level.