Inquiry Regarding the Calculation of Processing Units (PUs) with Batch Statistical API in Sentinel-hub

[Sagri]

I’m writing to you today to seek clarification regarding the calculation of Processing Units (PUs) when retrieving Sentinel-2 satellite data using your Batch Statistical API.

Firstly, I’d like to understand the PU calculation method when using a gpkg file that contains multiple polygons. Does the system compute the PUs per polygon and then aggregate them? For example, if we have four polygons (A to D), would the final PU calculation be an accumulation of the individual PUs from each polygon? (i.e., Total PU = PU of Polygon A + PU of Polygon B + PU of Polygon C + PU of Polygon D). Or, alternatively, does the system automatically create a bounding box (BBOX) that includes all polygons (A to D), and then calculates the PUs based on this encompassing polygon?

Secondly, I am curious about how the data acquisition period impacts the PUs. For instance, when retrieving data for a polygon of 512 * 512 pixels, if there’s only one day of observation, I understand that the Multiplication factor for the Area of Interest would be 1. But, if there are four days of observation, would the Multiplication factor increase to 4 (i.e., 512 * 512 pixels * 4 days)?

[gmilcinski (Grega Milcinski)]

Hi @Sagri,
for the first question - each polygon is counted separately
for the second - each observation counts, so in your case multiplier is indeed 4.

(as a side note - we have responded to this question via support e-mail as well a few days ago- might it be it got stuck in the spam?)

Best,
Grega

[Sagri]

Thank you for your response:)
I was able to confirm the definition, thank you.
(As you pointed out, the email had been classified as spam. I apologize.)