PM2.5 inversion method based on remote sensing and machine learning technology

• Main Authors: ZENG QINGHUAI, ZHANG ZHIGANG, XU WANG, YU XINFAN, XIONG XIANGYUN, YIN CHUNYANG, LIANG HONG, LIU KAI, YU LIANG, GONG LI, WANG WEIMIN, YU ZHAOKANG, WEN WEN, LI HUIYA, HE WEIBIAO
• Format: Patent
• Language: Chinese; English
• Published: 13.01.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIRCHEMICAL OR PHYSICAL PROPERTIES; MEASURING; PHYSICS; TESTING

The invention relates to the technical field of remote sensing, in particular to a PM2.5 inversion method based on remote sensing and machine learning technologies, which comprises the following steps: preprocessing meteorological station data, meteorological element data, topographic data, road network data and satellite remote sensing data; measuring and analyzing earth surface parameters, acquiring PM2.5 site data, analyzing spectral characteristics, and searching modeling characteristics; the method comprises the following steps: establishing a model, selecting and applying a model technology, training the model, setting and adjusting model hyper-parameters, verifying the model, developing and testing an integrated model, selecting an algorithm, optimizing the model, and combining PM2.5 site data to construct an inversion model of PM2.5; evaluating the performance of the model and establishing a reference, experimenting and adjusting the model in operation, and determining a machine learning model capable