A research team led by Dr. SHI Hailiang at the Anhui Institute of Optics and Fine Mechanics (AIOFM), Chinese Academy of Sciences, has developed a novel infrared imaging payload and AI-based retrieval framework capable of detecting carbon dioxide (CO₂) and methane (CH₄) emissions from space at a spatial resolution of approximately 100 meters. The system addresses key challenges in complex surface backgrounds and weak emission plume segmentation, enabling high-accuracy identification and quantification of point source greenhouse gas emissions.
These results have been published in International Journal of Applied Earth Observation and Geoinformation.
In response to the growing need for precise carbon emission inventories and verification, the team designed the “Hotspot Greenhouse Gas Monitoring Instrument” for an upcoming satellite mission. The payload features advanced imaging capabilities for CO₂ and CH₄ and represents a significant advancement in domestic high-resolution space-based GHG detection technology.
To overcome the difficulties caused by heterogeneous surface reflectance and high background variability—often leading to false positives or missed detections—the team proposed a Heterogeneous Surface Background Regulation Strategy, which utilizes multi-channel spectral fusion to significantly enhance detection performance over urban, desert, and high-aerosol regions. For accurate plume segmentation, the researchers developed two innovative methods: FSDINet, a dual-domain network architecture that integrates spectral features in both frequency and spatial domains for joint global-local modeling, and kMetha-Mamba, a hybrid approach that combines spectral derivative-based clustering with state-space modeling to improve weak plume detection and noise suppression.
This work lays a solid technical foundation for the data processing and emission retrieval applications of next-generation imaging carbon satellites in China. It marks a key step forward in high-resolution satellite-based monitoring of greenhouse gas emissions and contributes an innovative solution to global carbon tracking and scientific mitigation efforts.
Figure 1. Comparison of detection performance before and after signal interference suppression under different sXXurface background conditions. (Image by WU Shichao)
Figure 2. Comparison of performance improvements using different plume detection algorithms. (Image by WU Shichao)