【科技日报】新技术解决土壤水分遥感数据填补难题

Core Insights - A new technology framework that integrates machine learning and interpolation methods has been developed to address the common issue of large-scale data gaps in global satellite soil moisture remote sensing data products, significantly improving data completeness and usability [1][2] Group 1: Technology Development - The research team has proposed a novel approach that combines traditional interpolation methods and machine learning techniques to fill in missing soil moisture data [2] - The new method utilizes stacking heterogeneous ensemble technology to generate initial fill-in results from both interpolation and machine learning, followed by optimization through intelligent algorithms [2] Group 2: Performance and Applications - Experimental results indicate that the new technology performs excellently under various scales of data gaps, retaining the predictive capability of machine learning for large gaps while capturing local features through interpolation [2] - This technology is considered highly versatile and has the potential to be extended to the repair of various remote sensing data products, including surface temperature, vegetation parameters, and atmospheric components, providing higher quality data support for agriculture management, ecological protection, disaster monitoring, and climate change research [2]