search

qudt / qudt-public-repo

QUDT -Quantities, Units, Dimensions and dataTypes - public repository
Other
116 stars 73 forks source link

Pixel as a unit of length? #435

Open gwemon opened 3 years ago

gwemon commented 3 years ago

We have received a series of requests for new terms to be added to P01 vocabulary on the NERC Vocab Server for measurements of length, width and area of objects captured in digital images and expressed in units of "pixel". Can "pixel" be considered a valid unit of length?

steveraysteveray commented 3 years ago

This topic has cropped up in various places, with the general agreement that pixel is a "counting" unit rather than a unit of length. As evidence, multiplying a pixel width and a pixel height yields a pixel count - not pixel squared as it would be for a unit of length. Pixel is dimensionless. We could add it as such if that would be helpful.

jhodgesatmb commented 3 years ago

I have never heard of a pixel count. Everyone has heard of pixel rows and columns. Sadly, the size of a pixel is not constant so I am not sure it could be considered an area, but definitely not a length.

Jack Hodges, Ph.D. Arbor Studios

On Sep 20, 2021, at 9:17 AM, steveraysteveray @.***> wrote:

 This topic has cropped up in various places, with the general agreement that pixel is a "counting" unit rather than a unit of length. As evidence, multiplying a pixel width and a pixel height yields a pixel count - not pixel squared as it would be for a unit of length. Pixel is dimensionless. We could add it as such if that would be helpful.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

dr-shorthair commented 3 years ago

It could be argued for 'pixels' to be used as a unit of length or area. But not (generally) convertible to SI units, so outside the scope of QUDT (or P06), which is concerned with conversions.

jhodgesatmb commented 3 years ago

Certainly not outside the scope of QUDT.

Jack Hodges, Ph.D. Arbor Studios

On Sep 20, 2021, at 8:37 PM, Simon Cox @.***> wrote:

 It could be argued for 'pixels' to be used as a unit of length or area. But not (generally) convertible to SI units, so outside the scope of QUDT (or P06), which is concerned with conversions.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

gwemon commented 3 years ago

Thank you @steveraysteveray and @jhodgesatmb. So if I understand correctly, if we have a vocab that describes/identifies the variable that is being reported and a vocab that identifies the units, then the concept "pixel" should really be part of the variable description while the unit would need to be "dimensioneless" or "inapplicable"? I will think about it further and see how we could best handle this unsing our NVS P01 ("parameter labels") and P06 (units) vocabularies. Many thanks for your responses.

steveraysteveray commented 3 years ago

@gwemon, it looks like the group of us have still to reach perfect consensus, but my take is that a qudt:Quantity could be something like "my monitor's horizontal resolution", the qudt:Unit would be unit:Pixel, and the qudt:QuantityKind would be quantitykind:Count (which doesn't yet exist in QUDT), with a dimension vector of qkdv:A0E0L0I0M0H0T0D1 (dimensionless). The qudt:conversionMultiplier for unit:Pixel would be set to 0.0 which we have been using for cases where there is not a linear conversion to an SI unit.

We will need to discuss this more and come to a resolution. I agree with @jhodgesatmb that this is within scope, because QUDT is about more than just unit conversion.

justinprice02 commented 3 years ago

To add to @jhodgesatmb 's comment about adding pixel being within the scope of QUDT, it is conceivable that a pixel (or a type/types of pixel) does become connected to some SI unit in the future. Maybe this is through some industry standard or other convention. So having the unit and quantity kind would prepare QUDT to accommodate such developments for the purpose of conversions to SI units.

gwemon commented 3 years ago

Thank you all. It looks like I did close this ticket too early. In the meantime your feedback has been very useful and given me a way forward in modeling these new variables using our P01/P06 vocabularies.