New Deep Learning Approach to Boost Aerosol Measurement Accuracy in Space
by Simon Mansfield
Sydney, Australia (SPX) Feb 15, 2024
In a major development for atmospheric science, researchers have launched a novel algorithm designed to improve the accuracy of aerosol monitoring utilizing China’s FY-4A satellite tv for pc. This growth, detailed in a current publication in the journal Engineering, represents a collaborative effort among the many Institute of Atmospheric Physics (IAP) of the Chinese Academy of Sciences, the National Satellite Meteorological Center, the Harbin Institute of Technology, and a number of other different analysis institutes.
Aerosols, tiny particles suspended in Earth’s environment, play a vital function in understanding local weather change, air high quality, and the planet’s radiation steadiness. The FY-4A satellite tv for pc, geared up with the Advanced Geostationary Radiation Imager (AGRI), supplies important knowledge by scanning China each 5 minutes, enabling shut monitoring of those particles’ spatiotemporal variations.
Traditional strategies for measuring atmospheric aerosols have confronted challenges due to the rigidity of bodily retrieval algorithms and a scarcity of ground-based commentary websites. These limitations have hindered the applying of machine studying strategies, which require intensive knowledge samples for correct aerosol optical depth (AOD) retrieval.
To deal with these challenges, the analysis group developed an progressive algorithm that marries deep studying with switch studying strategies. This method attracts on ideas from each the darkish goal and deep blue algorithms, optimizing function choice for machine studying purposes in aerosol monitoring. By doing so, the algorithm considerably improves the pliability and accuracy of AOD retrieval from satellite tv for pc knowledge.
Independent validation of the algorithm has demonstrated its excessive accuracy in estimating AGRI aerosol ranges, showcasing a powerful correlation with anticipated values. This validation underscores the algorithm’s reliability and its potential as a predictive software for aerosol optical depth, marking a major step ahead in atmospheric monitoring from house.
Lead writer Fu Disong from the IAP emphasised the significance of this examine, stating, “Our examine showcases the numerous potential of merging the bodily method with deep studying in geoscientific evaluation.” Fu additional highlighted the broad applicability of their findings, noting, “The proposed algorithm holds promise for software to different multi-spectral sensors aboard geostationary satellites.”
This breakthrough not solely enhances the potential of China’s FY-4A satellite tv for pc to monitor atmospheric aerosols but in addition opens new avenues for making use of superior machine studying strategies in environmental and local weather science. By bettering the accuracy of aerosol measurements, scientists can achieve higher insights into atmospheric processes, aiding in the understanding and mitigation of local weather change and air high quality points.
Related Links
Institute of Atmospheric Physics (IAP) of the Chinese Academy of Sciences
Earth Observation News – Suppiliers, Technology and Application
https://www.spacedaily.com/reports/New_Deep_Learning_Approach_to_Boost_Aerosol_Measurement_Accuracy_in_Space_999.html
