Abstract
Abstract. Cloud coverage is one of the biggest concerns in spaceborne optical remote sensing, because it hampers a continuous monitoring of the Earth’s surface. Based on Google Earth Engine, a web- and cloud-based platform for the analysis and visualization of large-scale geospatial data, we present a fully automatic workflow to aggregate cloud-free Sentinel-2 images for user-defined areas of interest and time periods, which can be significantly shorter than the one-year time frames that are commonly used in other multi-temporal image aggregation approaches. We demonstrate the feasibility of our workflow for several cities spread around the globe and affected by different amounts of average cloud cover. The experimental results confirm that our results are better than the results achieved by standard approaches for cloud-free image aggregation.
Highlights
As determined by the MODIS mission, on average, about 67% of the Earth’s surface are covered by clouds (King et al, 2013), posing a well-known drawback for any remote sensing endeavours aiming at a monitoring of the Earth’s surface and relying on sensors operating in the optical domain
Google Earth Engine (GEE) is a web- and cloud-based platform for large-scale scientific analysis and visualization of geospatial data. It provides an extensive catalogue of remote sensing imagery and other geodata, as well as an application programming interface (API) with both JavaScript and Python front-ends allowing for the analysis of the data available in the catalogue on Google’s servers (Gorelick et al, 2017)
We made a Javascript version available via the GEE platform2. It consists of three main modules: (1) The Query Module for loading images from the catalogue, (2) the Quality Score Module for the calculation of a quality score for each image, and (3) the Image Merging Module for mosaicking of the selected images based on the meta-information generated in the preceding modules
Summary
As determined by the MODIS mission, on average, about 67% of the Earth’s surface are covered by clouds (King et al, 2013) (cf. Figure 1), posing a well-known drawback for any remote sensing endeavours aiming at a monitoring of the Earth’s surface and relying on sensors operating in the optical domain. In order to avoid the information gaps caused by clouds, Earth observation traditionally either resorts to sensors operating in the microwave domain or to algorithmic cloud removal strategies These are usually based on interpolation methods (Cihlar and Howarth, 1994, Zhu et al, 2012), machine-learning-based void filling approaches (Cheng et al, 2014, Chang et al, 2015, Huang et al, 2015, Xu et al, 2016), exploiting multi-sensor data fusion (Huang et al, 2015) or multi-temporal image sets (Lin et al, 2013, Cheng et al, 2014, Xu et al, 2016, Candra et al, 2017). ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume IV-2/W7, 2019 PIA19+MRSS19 – Photogrammetric Image Analysis & Munich Remote Sensing Symposium, 18–20 September 2019, Munich, Germany proposed method, before Section 5 summarizes and concludes our work
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
