[PROJ] How to write LAS points in N-E CRS

[Howard Butler]

> On Jun 29, 2020, at 8:26 AM, Javier Jimenez Shaw <j1 at jimenezshaw.com> wrote:
> 
> Hi
> 
> I am writing a LAS file using a Northing-Easting CRS (EPSG:2384 for instance).
> 
> I am not sure in which order I should write the point coordinates: northing-easting-up or easting-northing-up.
> * The first option (northing-easting-up) is "consistent" with the definition of the CRS. However it needs to understand the axes of the CRS (is this only possible in WKT2?), and operate the values to display them on the screen. Otherwise you will see a reflected image.
> * The second option (easting-northing-up) is much easier to represent in the screen. And seems to be what many programs are doing/expecting. However, it is inconsistent with the CRS definition.
> 

Javier,

Unfortunately there are not many satisfactory answers for you. The 1.4 LAS specification is deficient in that it says nothing about allowing any future versions of WKT, and it should be assumed that clients are only going to be able to consume WKT1:GDAL when writing data. I have attempted to start the wheels in motion to spin up the committee for another revision of the LAS specification to address this issue [1]. LAS 1.4 is going to be especially challenging in situations where orthometric data are written into time-dependent coordinates because there will be no way to communicate it. With the NGS delaying their rollout in the US, the immediate pressure caused by this need may have lessened, however.

I think the pragmatic answer to your question is to chose the latter approach, and then deal with the consequences once it is possible to more accurately model the axes in WKT2. 

Howard

[1] https://github.com/orgs/ASPRSorg/teams/lwg/discussions/16/comments/1