Abstract PROBA-V instrument launched in 2013 is offering a global daily coverage at pixel resolutions of 333 m and 1 km in three spectral bands (BLUE, RED, NIR) and 600 m for shortwave infrared (SWIR). The PROBA-V mission is the follow-on of the VEGETATION program started in 2000, which allowed generating long-term series at 1 km pixel resolution. The PROBA-V products belong to the Copernicus Global Land Service portfolio ( http://land.copernicus.eu/global/ ). The sensor design of PROBA-V with oriented cameras offers a wide field of view (FOV) for sampling the BRDF (Bidirectional Reflectance Distribution Function). This paper details the methodology implemented at the premises of VITO ( Flemish Institute for Technological Research ) with the aim to disseminate routinely from PROBA-V daily observations for both surface albedo (SA) and top-of-canopy corrected reflectance (TOC-R) products. The method classically operates a selection of cloudless scenes, performs atmospheric corrections, and finally applies a correction of directional effects on a pixel per pixel basis. The synthesis period is the decade and the composite period is 20 days. Such choice is a pointwise sampling as being a trade-off between the availability of clear scenes and the timescale for phenology. Regarding the albedo catalogue, a narrow-band to broadband conversion is stipulated. A recurrent technique serves for gap-filling based on the spread of weighed a priori data. Additional information concerns the quality flag and the age of the product. Preliminary accuracy assessment is performed through a comparison with the Moderate Imaging Spectroradiometer (MODIS) Collection 6. Dependable spatial consistency is reached except for wintertime with deviations in terms of rmse (root mean square errors) about 0.03 for visible and shortwave domains, and 0.04 for near infrared. Besides, both PROBA-V and MODIS C6 exhibit close time profiles, marked by smoothness or rapid transitions. Results over 10 confidence sites reveals rmse values of 0.032 and bias of 0.01 over the 2014 full annual cycle.

中文翻译：

---

哥白尼全球土地服务框架下PROBA-V仪器的表面反照率和toc-r 300 m产品

摘要 2013 年推出的 PROBA-V 仪器在三个光谱波段（蓝色、红色、近红外）和 600 m 的短波红外 (SWIR) 中以 333 m 和 1 km 的像素分辨率提供全球日常覆盖。PROBA-V 任务是 2000 年开始的 VEGETATION 计划的后续任务，该计划允许以 1 公里像素分辨率生成长期序列。PROBA-V 产品属于 Copernicus Global Land Service 产品组合 (http://land.copernicus.eu/global/)。带有定向相机的 PROBA-V 的传感器设计为 BRDF（双向反射分布函数）采样提供了宽视场 (FOV)。本文详细介绍了在 VITO（佛兰德技术研究所）场所实施的方法，目的是定期传播 PROBA-V 对地表反照率 (SA) 和冠层顶部校正反射率 (TOC-R) 的日常观测结果产品。该方法经典地操作选择的无云场景，执行大气校正，最后在每个像素的基础上应用方向效应的校正。综合期为十年，综合期为20天。这种选择是逐点采样，作为清晰场景的可用性和物候时间尺度之间的权衡。关于反照率目录，规定了窄带到宽带的转换。一种循环技术用于基于加权先验数据的传播来填补空白。附加信息涉及产品的质量标志和使用年限。通过与中等成像光谱仪 (MODIS) 集合 6 的比较来进行初步精度评估。除冬季外，可见和短波域的均方根误差（均方根误差）的偏差约为 0.03，达到了可靠的空间一致性。近红外。此外，PROBA-V 和 MODIS C6 都表现出接近的时间曲线，以平滑或快速过渡为标志。超过 10 个置信点的结果显示 2014 年整个年度周期的 rmse 值为 0.032，偏差为 0.01。除了冬季，可见和短波域的 rmse（均方根误差）偏差约为 0.03，近红外域的偏差约为 0.04。此外，PROBA-V 和 MODIS C6 都表现出接近的时间曲线，以平滑或快速过渡为标志。超过 10 个置信点的结果显示 2014 年整个年度周期的 rmse 值为 0.032，偏差为 0.01。除了冬季，可见和短波域的 rmse（均方根误差）偏差约为 0.03，近红外域的偏差约为 0.04。此外，PROBA-V 和 MODIS C6 都表现出接近的时间曲线，以平滑或快速过渡为标志。超过 10 个置信点的结果显示 2014 年整个年度周期的 rmse 值为 0.032，偏差为 0.01。