Does the converter automatically reprojects to EPSG: 4326?

[jpbolanoh]

Hi, I converter a .las pointcloud which has the crs 32618, but then when i try to visualize it on cesium it shows in a complete wrong location. I have been reading the documentation, but i could not find how mago3DTiler handles the crs.

Thanks Juan

[znkim]

Hello, @jpbolanoh

Yes, it will convert to the 4326 coordinate system by default.
Before that, though, it checks the coordinate system specified in the LAS header.
Sometimes the LAS format does not explicitly include the EPSG code.

In that case, you can try running it with the -crs 32618 command option.

Cheers

[jpbolanoh]

Hi, @znkim thanks for the quick answer. Yes, I add the -crs 32618 option when i convert the pointclouds but when i try to visualize it using cesium the tileset is located in a wrong place. I do not know what can be happening. If you want to see some of the files, feel free to ask.
Thanks
Quick update, the tileset is well placed when I convert a single pointcloud, when i ran the converter in a directory with several pointclouds, the tileset is located wrong. Now the problem that i face is that when i activate the WGS 84 ellipsoid the tileset is floating in the sky, and when i try with the cesium world terrain is placed below the ground. so, i want to know how the vertical datum is handler in the converter

[znkim]

I think this is an issue related to orthometric vs. ellipsoidal elevations.

Currently, mago-3d-tiler generates tiles based on the WGS84 ellipsoid.
So if your LAS files contain geoid-corrected (orthometric) elevations, you will likely see incorrect tileset heights in Cesium.

I’ll also look into the issue with converting multiple point clouds in a directory.
If you can provide a minimal sample to reproduce the problem, I’d be happy to help debug it.

[jpbolanoh]

Oh I see, so that height difference is due to the different vertical datum reference. So, these are the pointclouds that I have in my directory when i convert all the pointclouds at once. Also if you have time I would like to talk a little bit about the handler of the vertical datum.
https://drive.google.com/drive/folders/1VPASlgmoK5d1WkSqAMRdIPi6y5bqBGHY?usp=sharing
Thanks for the fast answers

[znkim]

I've checked the data and can summarize the situation as follows

First, Cesium World Terrain is based on ellipsoidal heights.
Meanwhile, your LAS files use orthometric(geoid-based) elevations.

As a result, when you visualize the LAS data on Cesium World Terrain, the tileset will appear lower offset by the local geoid height at that location.

To resolve this, you have two options

1. Visualize with terrain data that uses the same geoid reference, or
2. Convert your LAS files to use ellipsoidal heights before generating the tileset.

[jpbolano]

Hi @znkim, I tried to convert another data that i ask, and the person who generates told me that they used ellipsoidal heights to generate the laz file but still when i activate the Cesium World Terrain the point cloud appears below, like this

What else could be generating the error? If you want to check the laz file, is in the same url that i paste above in the directory WGS84 ELLIPSOID.
Sorry for ask a lot of questions