Handheld RGB-thermal images, and SAM segments for fine-scale heat exposure monitoring (Lusatia, Germany)

This dataset supports fine-scale monitoring of land-surface heat exposure using a low-cost handheld thermal camera.
It contains paired RGB and thermal imagery collected on a summer day across four landscape contexts in Lusatia, Germany (post-mining, urban, water-body, and agricultural sites). For each scene, we provide (i) the handheld images and (ii) segmentation outputs from the Segment Anything Model (SAM), including indexed mask maps that assign a segment ID to each pixel.
These files enable object- and land-surface temperature profiling and can be used for benchmarking segmentation/classification workflows or for microclimate analysis across land-surface types.