“The automated detection of methane emissions based on satellite remote sensing and artificial intelligence can provide technical support for methane emission control. The research team has constructed a methane plume dataset that considers complex spatiotemporal scenarios based on Sentinel-2 satellite data, proposed a deep learning detection model enhanced by spectral attention mechanism, and systematically evaluated the cross regional, temporal, and cross sensor transfer capabilities. It has achieved good application results in many places such as Central Asia and Africa, providing data and method support for intelligent monitoring of methane emissions through remote sensing.”