Questions about the coordinate reference system in Pix4Dmatic


I am a Pix4Dmapper user and I am now testing Pix4Dmatic. I have a few questions regarding the coordinate system.

  1. If I don’t use GCPs, selecting a coordinate system on “Select GCP coordinate system” would convert my current system to a diffent one? Is it similar to the output coordinate we selected on Pix4Dmapper?

  2. If I do use GCPs, whatever coordinate system I select on the GCP tab would become the project’s output coordinate system, right? It would override the default image’s coordinate system in this case?

  3. Is there a geoid model for Brazil like Mapgeo2015 or hgeoHNOR2020? If not, how do I set the height over a specif geoid, please?

  4. Last question, the altitude presented as ellipsoidal height when I uploaded aerial images like the urban area from the Pix4Dmatic example dataset. Is that altitude MSL capture by the drones’s barometer or is that the altitude over the ellipsoid WGS84 for example?

I am loving the software and I appreciate any help,

Thank you for your time.

Hi @rfbessa

My replies below:

Yes, it’s a little different than in PIX4Dmapper, but changing the GCP coordinate system modifies what we call the “project coordinate system” in PIX4Dmatic. The latter is similar to what you know as “output coordinate system”. The only thing to notice is that in the case that you select a geographic coordinate system for this GCP coordinate system, the project coordinate system is still going to be a projected coordinate system. Otherwise, it’s going to be the same.

Yes, see my reply to 1).

Not for the moment. My understanding is that we’re missing a vertical coordinate reference system in the EPSG catalog. Not sure if you have links to the Brazilian geodetic authorities? But this may be interesting to ask them. If they add it to the EPSG, we should be able to support this relatively easily. You can find more documentation on setting the geoid height here:

Not 100% sure, but it’s probably altitude over the ellipsoid.

Hope you’ll enjoy using PIX4Dmatic! Don’t hesitate if you have more questions :wink:

Thank you for answering my questions:

I do have the link in English for the geodetic authority in Brazil, it is the link down below:

I have 3 other questions if you don’t mind:

  1. When I select ellipsoidal height over the GRS 1980 ellipsoid (image 1 below), the software maintains the MSL height recorded by the drone, correct? Or should I select the EPSG 5714 MSL height?

  1. Since there is not a geoid model for Brazil, If I select the EGM96 geoid for example (image 2 below), I can select the height of the Brazilian Geoid over the EGM96 like this, correct?

  1. What does the polygon hole tool do in Pix4Dmatic? What does the backface culling and object occlusion settings do? I’ve been trying to test it or find a detailed explanation or video but couldn’t.

Thank you for your time.

Okay, here we go:

Thanks! Have you written them to ask about it? It’s probably good if Brazilians write them directly. We can try too on our side.

The “input CRS” (it’s the CRS of the images, or cameras) and the “project CRS” are not the same, i.e. the drone data is “read” by the software with its associated CRS. You can see this in the cameras tab under each camera model. There you can also change the CRS of the images if you need to for some reason (btw on the “internals” button on the right of each camera, you can see the internal parameters too) :

This means that whatever you select for the project CRS is what you want your project to be in, the software will then transform the input data from the input CRS to your project CRS.


For example, if you model a house with the (connected) polygons but there is a chimney on the roof. You’d make a polygon for one side of the roof, then you could make a hole where the chimney is and create more polygons to model the chimney correctly. All these polygons can be connected AND there can be a “planar” constraint on the polygons too. This will optimize the entire surface together (due to the connectivity) AND it will make sure that faces are planar.

The backface culling is in case you have an indoor scene for example, it’s hiding everything that’s not facing you, e.g. if you have a mesh and you look at it from the back, you wouldn’t see that part anymore when the backface culling is enabled. (hence the “backface” and I suppose the “culling” is for removing when you see the backface).

Object occlusion is about when you see an object “through” other elements in the scene.

Hope that helps!