A decision support tool for the optimization of compositing parameters

Abstract

Temporal syntheses of surface reflectance are one of the most common data products from high temporal resolution instruments. Such an image combination procedure is sensitive to several control parameters, including the compositing period. Unfortunately, these choices usually rely on a unique and global solution delivered by the data provider to ensure temporal and spatial consistency across applications. However, many applications require customized composites to achieve their goals and meet regional constraints. Although this need is now widely recognized by the scientific community, practitioners still rely on trial and error and ad hoc adjustments in designing their approach. The objective of this paper is to design, implement, and test a decision support tool that can find the most appropriate compositing parameters for coarse‐ to medium‐resolution sensors. The key innovation of this approach is that it incorporates external data on the cloud cover and seasonality of the region studied. The algorithm can be applied to any region of the globe and to any optical satellite instrument recording surface reflectance over time. The tool allows data users to optimize the compositing parameters to their application subject and to regional conditions. It can also be used to determine the feasibility of a proposed compositing process. The potential of this methodology to improve the relevancy and the quality of time series products is demonstrated by testing it on two specific applications.

Acknowledgements

The authors are grateful to the Joint Research Center for providing SPOT VGT data from the year 2000, in the framework of the Global Land Cover 2000 program. They are also grateful to the VGT Programme for providing VGT Images from 2001 to 2004. They also would like to acknowledge the financial support of the Belgian National Fund for Scientific Research (FNRS) by way of a FRIA grant. Finally, they thank Baudouin Desclee, Bruno Gerard and Gregory Duveiller for their relevant comments.