河海大学学报（自然科学版）2025，Vol.53Issue(6)：75-81,109,8.DOI:10.3876/j.issn.1000-1980.2025.06.009
基于微波遥感和多模型集成的安徽省土壤湿度反演
Soil moisture inversion in Anhui Province based on microwave remote sensing and multi-model ensemble
摘要
Abstract
To accurately invert the soil moisture in Anhui Province and enhance the adaptability and accuracy of machine learning models,support vector regression,extreme gradient boosting,CatBoost,random forest,AdaBoost,and Stacking model(the first five machine learning models were selected as the base models,linear regression was used as the meta-model)were used to invert the soil moisture in Anhui Province.The spatial distribution and time series of soil moisture inversion results obtained from the Stacking model were analyzed.The results indicate that,with the same input remote sensing data,the Stacking model has higher accuracy and robustness compared to individual models.The correlation coefficient between the inverted soil moisture and the measured data reaches 0.72,and the root mean square error(RMSE)is 0.05 m3/m3.The spatial heterogeneity of soil moisture in Anhui Province is relatively high.The northern region is relatively dry,with an average soil moisture of around 0.2 m3/m3.The eastern areas of Chaohu Lake and the Yangtze River region are relatively humid,with an average soil moisture of up to 0.4 m3/m3.Although the Dabie Mountain area and the southern part of Anhui Province have higher altitudes,their soil moisture is still higher than that of the Huaibei Plain,indicating that the differences in climate between the north and the south may affect the magnitude of soil moisture.Overall,the spatial pattern of soil moisture in Anhui province is an increasing trend from the northwest to the southeast.关键词
土壤湿度/集成学习/遥感反演/机器学习/风云卫星/安徽省Key words
soil moisture/ensemble learning/remote sensing inversion/machine learning/Fengyun Satellite/Anhui Province分类
信息技术与安全科学引用本文复制引用
孔月,冯平,张珂,申晓骥,王晟,王宇昊,田鑫丽,何蒙,刘匡,向征..基于微波遥感和多模型集成的安徽省土壤湿度反演[J].河海大学学报（自然科学版）,2025,53(6):75-81,109,8.基金项目
国家重点研发计划项目(2023YFC3006505) （2023YFC3006505）
中央高校基本科研业务费专项资金项目(B240203007) （B240203007）
国家自然科学基金项目(52309017) （52309017）
江苏省自然科学基金项目(BK20230958) （BK20230958）
