One of the members of the NICFI Satellite Data Program asks for support on how to run code, at scale, in Google Earth Engine. If you have experienced and solved a similar problem, let us know if you are able to help! You can also check out this thread.
“Currently, I am engaged in a significant endeavor to map forest cover. I am utilizing images from the multispectral Planet mosaics as part of the NICFI initiative, which has proven to be a valuable resource for this activity. However, I am facing significant challenges in processing this data at a national scale, particularly in terms of excessive resource consumption, such as memory, when conducting large-scale analyses (800k sqkm) on the GEE platform.
Within the scope of this project, I have been trying to divide the national surface area down to the provincial level to address this issue. Unfortunately, I am still encountering considerable memory overruns. I would like to inquire whether you might have any suggestions for optimizing the code and developing effective strategies to overcome these memory challenges within the GEE environment, or if there is a possibility to recommend someone with expertise in the GEE platform or open-source tools and codes for processing.
Additionally, I am interested in understanding the best recommended practices for processing large volumes of Planet data and effective methods for cloud and shadow removal in this data within the GEE environment.”
Let us know if you are able to help by commenting below!
