Datasets › TBBR
TBBR (Thermal Bridges on Building Rooftops)
The dataset of Thermal Bridges on Building Rooftops (TBBR dataset) consists of annotated combined RGB and thermal drone images with a height map. All images were converted to a uniform format of 3000×4000 pixels, aligned, and cropped to 2400×3400 to remove empty borders.
The raw images for our dataset were recorded with a normal (RGB) and a FLIR-XT2 (thermal) camera on a DJI M600 drone. They show six large building blocks of around 20 buildings per block recorded in the city centre of the German city Karlsruhe east of the market square. Because of a high overlap rate of the images, the same buildings are on average recorded from different angles in different images about 20 times.
All images were recorded during a drone flight on March 19, 2019 from 7 a.m. to 8 a.m. At this time, temperatures were between 3.78° C and 4.97° C, humidity between 80% and 98%. There was no rain on the day of the flight, but there was 2.3 mm/m² 48 hours beforehand. For recording the thermographic images an emissivity of 1.0 was set. The global radiation during this period was between 38.59 W / m² and 120.86 W / m². No direct sunlight can be seen visually on any of the recordings.
The dataset contains 924 images with a total of 6930 annotations of thermal bridges on rooftops, split into train and test subsets with 722 (5614) and 202 (1313) images (annotations), respectively. The annotations only include thermal bridges that are visually identifiable with the human eye. Because of the aforementioned image overlap, each thermal bridge is annotated multiple times from different angles.
For the annotation of the thermal images the image processing program VGG Image Annotator from the Visual Geometry Group, version 2.0.10, was used. The thermal bridge annotations are outlined with polygon shapes. These polygon lines were placed as close as possible but outside the area of significant temperature increase. If a detected thermal bridge was partially covered by another building component located in the foreground, the thermal bridge was also marked across the covering in case of minor coverings. Adjacent thermal bridges, which affect different rooftop components, were annotated separately. For example, a window with poor insulation of the window reveal located in the area of a poorly insulated roof is annotated individually. There is no overlap between annotated areas. While each image contains annotations, they also include thermal bridges present that are not annotated.
Usage:
Each compressed archive file represents one of the six building blocks. For the related publication the final block (Flug1_105Media) was used as a hold-out test sample. The archives contain Numpy files (one per image) of shape (2400, 3200, 5), where the final dimension is the color channel in the format [B, G, R, Thermal, Height].
Archives were compressed using ZStandard compression. They can be decompressed in a terminal by running e.g.
tar -I zstd -xvf Flug1_105Media.tar.zst
these will be decompressed into the file structure:
images/
└── Flug1_105Media/
└── DJI_0004_R.npy
└── DJI_0006_R.npy
└── ...
Corresponding annotations are provided in the COCO JSON format. There is one file for training (Flug1_100Media - Flug1_104Media blocks) and one for test (Flug1_105Media block). They contain a single class (thermal bridge) and expect the folder structure shown below.
Note: The annotation files contain relative paths to numpy files, in case of problems please convert to absolute paths (i.e. insert the containing directory before each file path in the JSON annotation files).
We recommend the following folder structure for reproduction of our work with Detectron2:
├── train/
│ ├── Flug1_100-104Media_coco.json
│ └── images/
│ ├── Flug1_100Media/
│ │ ├── DJI_XXXX_R.npy
│ │ └── ...
│ ├── ...
│ └── Flug1_104Media/
│ ├── DJI_XXXX_R.npy
│ └── ...
└── test/
├── Flug1_105Media_coco.json
└── images/
└── Flug1_105Media/
├── DJI_XXXX_R.npy
└── ...
Benchmarks archive 2025-07-28
All 2 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.
| First row (archive order) | Paper | Code | ||||
|---|---|---|---|---|---|---|
| Object Detection | TBBR | Swin-T (ImageNet-1k pretrain) Average Recall@IoU:0.5-0.95 45.4 | Deep learning approaches to building rooftop thermal... | Helmholtz-AI-Energy/TBBRDet | 7 | Compare |
| Instance Segmentation | TBBR | Swin-T (ImageNet-1k pretrain) Average Recall@IoU:0.5-0.95 28.0 | Deep learning approaches to building rooftop thermal... | Helmholtz-AI-Energy/TBBRDet | 5 | Compare |
Papers archive 2025-07-28
papers with a benchmark row: 2 · with a code link: 1 · where Syntology ran a sample: 0 (0 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument) Syntology
Show: all papers with a benchmark rowonly where code ran: none of the 2 papers with a benchmark row has a Syntology run on record
Page 1 of 1: papers 1 to 2 of 2 with a leaderboard row on this dataset's benchmarks, newest first by the archive's date (ties by slug; undated papers last).
The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset, not that list; the archive's count for this dataset is 2. The Syntology column is from Syntology's graph, stated per sample; it is not part of any archive number. A line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified” (C is a part of N, never taken away from it); the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced. When the archive marks a repository official for the paper, the cell starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record.
| Paper | Code | Results | Date | Samples run Syntology |
|---|---|---|---|---|
| Deep learning approaches to building rooftop thermal bridge detection from aerial images | 1 | 11 | 12 Dec 2022 | not harvested |
| AI-based thermal bridge detection of building rooftops on district scale using aerial images | 0 | 1 | 9 Aug 2021 | not harvested |
Dataset loaders archive 2025-07-28
No loader listed in the archive.
Tasks archive 2025-07-28
License archive 2025-07-28
Creative Commons Attribution 4.0 International
Modalities archive 2025-07-28
Languages archive 2025-07-28
Variants archive 2025-07-28
- TBBR
1 variant name, as the archive lists them.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections