VOC-IRLeak: Infrared Video Dataset for VOCs Gas Leakage Detection

VOC-IRLeak Dataset v1.0 📋 Dataset Overview VOC-IRLeak is an infrared video dataset designed for VOCs (Volatile Organic Compounds) gas leakage detection. This dataset provides raw infrared video sequences for training and evaluating gas leakage detection and source localization algorithms. Purpose This dataset is designed to support research and development in: Gas leakage detection using infrared imaging Source localization of gas leaks Anomaly detection in industrial environments Computer vision applications for environmental monitoring Industrial safety monitoring systems Data Collection Acquisition Equipment: Model: GL1000i infrared camera Working Band: 3.2 - 3.5 μm (mid-wave infrared) Detector Array: 320×256 pixels (standard cooled detector) Output Resolution: 1920×1080 pixels Frame Rate: 30 frames per second (fps) Video Format: MP4 Collection Environment: Scenarios: Industrial and experimental environments Gas Type: Methane (CH₄) with composition of 99% CH₄ and 1% H₂ Applicable Gas Types: This dataset is primarily focused on methane detection, but the methodology and data format are applicable to other VOCs that exhibit absorption characteristics in the 3.2-3.5 μm wavelength range Leakage Conditions: Various leakage scenarios including different flow rates, distances, and environmental conditions Dataset Characteristics Video Format: MP4 Resolution: 1920×1080 pixels Frame Rate: 30 fps Content: Raw infrared video sequences showing gas leakage events ⚠️ Citation Requirement If you use this dataset in your research, you MUST cite both the dataset and the associated paper: 📌 How to Cite in Your Paper In your paper's reference section, please include BOTH citations: Primary Citation (Paper) - Cite the associated paper in your main references: This is the primary citation that should appear in your paper's reference list DOI: 10.1016/j.measurement.2025.119085 Dataset Citation - Also cite the dataset (can be in data availability section or footnotes): DOI: 10.5281/zenodo.18780694 1. Paper Citation (Primary) @article{shen2026source, author = {Shen, H. and Xu, L. and Dong, S. and Xu, Q. and Yu, H. and Wang, M. and Li, F.}, title = {Source localization of infrared gas leakage based on YOLOv8 and Gaussian diffusion models}, journal = {Measurement}, volume = {258}, pages = {119085}, year = {2026}, doi = {10.1016/j.measurement.2025.119085}, url = {https://doi.org/10.1016/j.measurement.2025.119085} } Plain text format: Shen, H., Xu, L., Dong, S., Xu, Q., Yu, H., Wang, M., & Li, F. (2026). Source localization of infrared gas leakage based on YOLOv8 and Gaussian diffusion models. Measurement, 258, 119085. https://doi.org/10.1016/j.measurement.2025.119085 2. Dataset Citation @dataset{shen2026vocirleak, author = {Shen, H. and Xu, L. and Dong, S. and Xu, Q. and Yu, H. and Wang, M. and Li, F.}, title = {VOC-IRLeak: Infrared Video Dataset for VOCs Gas Leakage Detection}, year = {2026}, publisher = {Zenodo}, doi = {10.5281/zenodo.18780694}, url = {https://doi.org/10.5281/zenodo.18780694} } Plain text format: Shen, H., Xu, L., Dong, S., Xu, Q., Yu, H., Wang, M., & Li, F. (2026). VOC-IRLeak: Infrared Video Dataset for VOCs Gas Leakage Detection [Data set]. Zenodo. https://doi.org/10.5281/zenodo.18780694 Important Notes: The paper citation (DOI: 10.1016/j.measurement.2025.119085) should be included in your paper's main reference list The dataset citation (DOI: 10.5281/zenodo.18780694) should also be included, typically in a "Data Availability" section or as a footnote Failure to cite both the dataset and the associated paper constitutes academic misconduct 📊 Dataset Information Version: 1.0 License: CC BY 4.0 (Creative Commons Attribution 4.0 International) Format: Raw infrared video sequences (MP4 format) Purpose: VOCs gas leakage detection and source localization Dataset DOI: https://doi.org/10.5281/zenodo.18780694 ⚠️ Important Notice on Dataset Scope Due to equipment copyright and other restrictions, this dataset represents only a subset of the original dataset used in the associated paper. The dataset may be updated periodically in the future. Please check the GitHub repository or Zenodo page for the latest version. 📝 Segmentation Annotation Availability Segmentation annotation data may be provided in the future. If you need segmentation annotations for your research, please contact the authors via email: [e-mail] We will consider providing annotation data based on research needs and availability.