SentinelKilnDB
A Large-Scale Dataset and Benchmark for OBB Brick Kiln Detection in South Asia Using Satellite Imagery
NeurIPS 2025 Satellite Imagery Object Detection Environmental Monitoring
62,671 Brick Kilns
2.8M km² Coverage
4 Countries
3 Kiln Types
Overview
Air pollution was responsible for 2.6 million deaths across South Asia in 2021 alone, with brick manufacturing contributing significantly to this burden. The Indo-Gangetic Plain sees brick kilns contributing 8-14% of ambient air pollution.
SentinelKilnDB addresses the critical need for automated brick kiln detection by providing the first publicly available, hand-validated benchmark of 62,671 brick kilns with oriented bounding boxes (OBBs) across three kiln types using free Sentinel-2 imagery.
Dataset Statistics
| Country | CFCBK | FCBK | Zigzag | Total | Coverage |
|---|---|---|---|---|---|
| India | 1,939 | 21,451 | 19,592 | 42,982 | 9 states |
| Bangladesh | 2 | 1,461 | 5,440 | 6,903 | 8 divisions |
| Pakistan | 3 | 10,443 | 1,731 | 12,177 | 4 provinces |
| Afghanistan | 0 | 608 | 1 | 609 | 34 provinces |
| Total | 1,944 | 33,963 | 26,764 | 62,671 | 2.8M km² |
Kiln Types
CFCBK Kilns
Circular Fixed Chimney Bull’s Trench Kilns - oldest design, fuel-intensive, high emissions
FCBK Kilns
Fixed Chimney Bull’s Trench Kilns - most prevalent (70-75%), newer design, more efficient
Zigzag Kilns
Most efficient design with 40% fuel savings compared to FCBK kilns (20-25% of total)
Abstract
Air pollution was responsible for 2.6 million deaths across South Asia in 2021 alone, with brick manufacturing contributing significantly to this burden. In particular, the Indo-Gangetic Plain; a densely populated and highly polluted region spanning northern India, Pakistan, Bangladesh, and parts of Afghanistan sees brick kilns contributing 8-14% of ambient air pollution.
Traditional monitoring approaches, such as field surveys and manual annotation using tools like Google Earth Pro, are time and labor-intensive. Prior ML-based efforts for automated detection have relied on costly high-resolution commercial imagery and non-public datasets, limiting reproducibility and scalability.
In this work, we introduce SentinelKilnDB, a publicly available, hand-validated benchmark of 62,671 brick kilns spanning three kiln types Fixed Chimney Bull’s Trench Kiln (FCBK), Circular FCBK (CFCBK), and Zigzag kilns—annotated with oriented bounding boxes (OBBs) across 2.8 million km² using free and globally accessible Sentinel-2 imagery.
We benchmark state-of-the-art oriented object detection models and evaluate generalization across in-region, out-of-region, and super-resolution settings. SentinelKilnDB enables rigorous evaluation of geospatial generalization and robustness for low-resolution object detection, and provides a new testbed for ML models addressing real-world environmental and remote sensing challenges at a continental scale.
Citation
@article{mondal2025sentinelkilndb,
title={SentinelKilnDB: A Large-Scale Dataset and Benchmark for OBB Brick Kiln Detection in South Asia Using Satellite Imagery},
author={Mondal, Rishabh and Parab, Jeet and Kubadia, Heer and Dubey, Shataxi and Junagade, Shardul and Patel, Zeel B and Batra, Nipun},
journal={Advances in Neural Information Processing Systems},
year={2025}
}