Deep transfer learning assisted detection of methane super-emitters in oil and gas fields using Sentinel-2 observations

In the folder you will find the six methane super-emitters retrieval (ΔR) imagery datasets and the synthetic ΔR datasets described in the article "A Data-Efficient Deep Transfer Learning Framework for Methane Super-Emitter Detection in Oil and Gas Fields Using Sentinel-2 Satellite" Zhao et al. (2025). The dataset files are provided in .png format. The name of each file indicates: -the methane source index (see Table 1): D1 = Dataset#1-the acquisition date All images were retrieved from 20m Sentinel-2 L1C products and the retrieving domain size is 16 km2.

It is noted that differences in the methane point source retrieval method workflow and the analyst's labeling rules may lead to variations in the resulting methane plume images. If any omissions or errors are found in the dataset during use, please contact the authors for corrections. Importantly, if this dataset is used for publication, please cite the related article.