Practice and Reflection on the Construction of Remote Sensing Products Validation Network for the Land Observation Satellite
MA Lingling,ZHOU Xiang,WANG Ning,TAO Zui,ZHAO Yongguang,ZHU Xiaohua,GAO Caixia,YANG Jian,GAO Hailiang,YANG Hong,XIAO Qing,LI Qiangzi,LYU Tingting,ZHANG Fengli,ZANG Wenqian
DOI: https://doi.org/10.11834/jrs.20244100
2024-01-01
Abstract:Validation of remote sensing products,as a crucial process bridging remote sensing data products and their application services,is essential for meeting the increasing demands for precision and performance in a series of quantitative applications.Such a validation is also necessary for improving the algorithm and production procedure after collecting feedback in such applications. Over the years,numerous sites for the validation of land remote sensing products have been established domestically and internationally.However,the theoretical and technical systems in the construction of a validation network is still not perfect in the entire process.These systems include the site characteristic representation and selection,measurement of surface and/or atmospheric parameters,sampling of measurement,and validation service mode toward different users.The possible reasons include the complexity and dynamic nature of the Earth system that lead the interaction between different spheres,land surface heterogeneities and uniformities,system and random errors in measurements,and various types of remote sensing satellite products for different applications.These problems result in difficulties in fully utilizing existing facilities.Therefore,to solve a series of basic theories and techniques problems,and to provide new models and specific solutions become the precondition in improving the application efficiency of validation system infrastructure at present and in the near future. In this research context,the status and development trends are first analyzed,including the research domain of ground observation networks for validation,acquisition technology for ground reference truth,and comprehensive validation services for remote sensing products domestically and internationally.Thereafter,three basis theories(i.e.,spatial variability theory in geostatistics,uncertainty theory in metrology,and optimization theory in operations research)are adopted to improve the validation methods.These three theories explore the spatial representativeness characterization of the network of product validation,uncertainty analysis and transfer of the fiducial reference measurements,comprehensive weighting and composition of multiple validation results,and optimal balance between the validation resources and user demands.Based on the theories,a framework is developed for the methodological system of the terrestrial observation satellite product validation network.This framework lays the theoretical foundation for the construction of a remote sensing production validation system and plays an important supporting role in forming targeted solutions.Aiming with the construction of the National Civil Space Infrastructure and the application needs of validation of key common products,considering overall layout,product coverage completeness,spatio-temporal consistency,and traceability,this study proposes specific requirements for the validation targets,validation areas,validation methods,validation accuracy,validation frequency,and service mode of Chinese land observation satellite remote sensing product validation network.The blueprint of the validation network for the"14th Five Year Plan"was designed based on the proposed methodology. This study coordinates the construction of a comprehensive product validation network,comprising an aerial collaborative validation system,a benchmark referenced transfer and fiducial reference measurement system,and a high accuracy validation service system.This network forms a technical solution for Chinese land observation satellite remote sensing product validation.The proposed solution will promote the resolution of related bottleneck issues,such as the accuracy,efficiency,and consistency of product validation.Lastly,the proposed solution will significantly improve the theoretical method system of the product validation network by fully utilizing and optimizing the use of domestic land observation satellite remote sensing data,and enhancing the level of quantitative applications.