Menu
Terrascope logo Proba-V MEP logo
This forum applies to Terrascope and Proba-V MEP. Please note that the forum is in English only.
Please log in using the link in the top menu to add new forum topics or comments.

Sentinel 2 image size

Submitted by p3tzl on
Hello everyone, I'm using Terrascope's jupyterlab in order to analyze the NDVI, NDBI and NDWI of a certain area over a time range. While analyzing the selected files, I noticed that while I selected the same AOI everywhere, I don't get the same image shape every time and the area observed in my image is not the same. To make my problem clearer, I have attached my code here: https://we.tl/t-jtpM5Is9W8   In the last cell, the different shapes of the considered files are printed. Do I have to isolate my AOI in another way, because I thought that using mask.msk only the aoi is considered?
Tags
Coding

Comments

Submitted by daemsd on Thu, 16/11/2023 - 16:13

Hi, this is expected because not all files have the same intersection with your AOI. When looking at the Sentinel-2 tiling grid, you will notice that several tile IDs (e.g. 31UES, 31UER, ...) intersect with your AOI; data from these tile IDs have a different intersection with your AOI, which results in different shapes. Regards, Dirk

Submitted by p3tzl on Thu, 16/11/2023 - 21:32

Hello Dirk, thank you for the reply! I assume this is because the AOI is chosen so big? I found that I have to fiddle around a bit with the AOI in order to make all tile IDs intersect similarly

Submitted by driesj on Fri, 17/11/2023 - 07:18

Hi, just a note: for this kind of problem, Terrascope offers the openEO API via openEO platform: https://openeo.cloud/ This generally solves all difficulties from trying to work directly with the complex file based products. Computing a timeseries for such indices is for instance a common use case.
stay up to date!
subscribeto our newsletter
new perspectives

Stay up to date!