Sensing time at pixel level

Hi everyone,

From the Sentinel-2 Products Specification Document (https://sentiwiki.copernicus.eu/__attachments/1692737/S2-PDGS-CS-DI-PSD%20-%20S2%20Product%20Specification%20Document%202024%20-%2015.0.pdf?inst-v=e48c493c-f3ee-4a19-8673-f60058308b2a) I understood that the sensing time for a granule/tile is the average of all individual pixels. Now if I understand correctly, the sensing time of a specific pixel is based on which detector captured that part of the granule/tile.

How do I get the exact time that a pixel has been captured in the final image?

In the DATASTRIP meta data I found that there are GPS_times for a REFERENCE_LINE. Can I use this to calculate the time offset between the average sensing time of the granule/tile and the exact time of the pixel?

Kind regards,

Thymen