全球长时间序列遥感植被指数产品去云重建算法研究

Long-term remotely sensed vegetation index data set is vital for the study of global vegetation dynamic. The less than ideal atmospheric conditions such as cloud and aerosol result in the spatiotemporal discontinuity in these data set, which impedes the relevant application of global vegetation monitoring. The time series reconstruction can be applied to eliminate the effect of discontinuous problem. With numerous studies on the time series reconstruction in remote sensing, decades of reconstruction models had been applied to various applications, which powerfully promoted the development of downstream application based on vegetation index data set. Harmonic Analysis of Time Series (HANTS) is one of the most widely used reconstruction models. Considering the lack of fine-performance global reconstruction scheme and the unknown effect of reconstruction processing on downstream applications, this project plan to systematically evaluate the impact of some uncertain factors of HANTS on global reconstruction performance. Then a global reconstruction scheme based on the improved HANTS will be developed. Further, we will quantify the impact of reconstruction processing on vegetation index based applications such as phenology, drought, and vegetation trending monitoring. This study can effectively improve the performance of global reconstruction, which will provide much more powerful data support for long-tern vegetation index based global change research.

长时间序列遥感植被指数产品是研究全球陆表植被变化的重要数据支撑。云和气溶胶等不理想大气条件造成的时空不连续问题阻碍了相关的全球植被监测应用。时间序列重建处理可以有效地消除上述时空不连续问题的影响，学界已经在遥感时间序列重建领域开展了大量的研究，开发了数十种重建模型及方案，有效地推动了基于遥感植被指数产品的应用研究进展。HANTS（Harmonic Analysis of Time Series）模型是最为常用的重建模型之一。针对当前缺乏高精度全球重建方案以及重建处理对下游应用的影响不明的问题，本研究拟通过系统评价HANTS重建过程中一些不确定因素的全球重建的影响，发展一套改进的HANTS全球重建方案。还将界定重建处理对基于植被指数产品开展的诸如物候、干旱及植被趋势监测等应用的影响。本项目的顺利开展将大幅提高当前全球重建模型的精度，为基于长时遥感数据产品的全球变化研究提供更有力的数据支撑。

云和气溶胶等不理想的大气条件导致了大量基于光学（可见光、见红外、热红外）波段反演得到的陆表遥感产品（如NDVI、LAI和LST）呈现出时空不连续的特点，阻碍了这些产品的广泛应用。对长时序陆表遥感产品进行时空重建成为解决这一问题的重要手段。时间序列谐波分析（HANTS）算法是数十年来最为常用的一种遥感时间序列重建算法。本项目系统评估了HANTS算法在全球重建中的精度，提出了优化的全球重建参数设置，得到了全球优化HANTS模型。此外，项目还基于云计算平台实现了可拓展的HANTS时空重建模块，用于支持多元化用户可定制的全球遥感数据时空重建处理业务。本项目的成果将经验主导的区域HANTS模型应用推向了系统优化后的全球重建应用，且通过模块化的方式使得模型能够更好的响应非专业用户的重建处理需求，进一步推进HANTS算法的普适化应用。