Cleanup the 3d data from the photogrammetry tiles¶

When the photogrammetry tiles have been optimized, it is now possible to cleanup their 3d data, i.e. remove 3d data that can interfere with the MSFS generated 3d assets, mainly the water area and the vegetation area.

All you have to do is to select Google Earth Decoder Optimization Tools > 4. Cleanup 3d data fro Google Earth tiles in the top menu.

Set up the cleanup of the 3d data¶

First of all, you have to decide the method you will use to cleanup the 3d data. Those methods rely on the accuracy of the OpenStreetMap data. Two levels of accuracy are available: Accurate and Not accurate.

Check for OpenStreetMap data accuracy¶

If you indicate to GEDOT that the OpenStreetMap (OSM) 3d data are accurate, GEDOT will use a constructive method to cleanup the data: this means that GEDOT will keep only the relevant data from the tiles (for instance, buildings and roads).
If you indicate to GEDOT that the OpenStreetMap (OSM) 3d data are not accurate, GEDOT will use a destructive method to cleanup the data: starting from the complete tile, it will remove the non relevant 3d data (for instance, forests, woods, sea, lake, ...)

Why does the accuracy of OSM data determine the method that GEDOT uses to cleanup the 3d data ?

To be able to keep only the relevant 3d data (constructive method), GEDOT needs to know the shape of the buidlings in OpenStreetMap.
When the OSM data are accurate, this means that almost all the buidlings (more than 95%) are represented in the OpenStreetMap area.

If it is not the case (remember that OpenStreetMap is an open source project, based on the users' contributions), GEDOT cannot know the 3d data to keep. So the logic behind needs to be different. In this case, GEDOT uses a destructive method to remove the area that it can retrieve from the OSM data (water, forests, woods, etc...).

How can i check for OpenStreetMap data accuracy ?

Open your Internet browser, and just ask this on the research bar: "<Your town> OSM relation" (for instance, "Arcachon OSM relation"). Click on the first link proposed (this should be "Relation: Your town (########)", for instance "Relation: Arcachon (1665459)")

An example is provided here:

Now, you can check for the accuracy of the OpenStreetMap data, and especially if the building shapes are mostly all available (in the case of Arcachon city, the 3d data are accurate, because they rely on cadastral data).

Below is an illustration of a city with accurate OpenStreetMap data (Rotterdam):

Below is an illustration of a city with not accurate OpenStreetMap data (Rio de Janeiro):

You see the difference ? In the Rotterdam 3d data, all the buildings are drawn (dark grey shapes), which is not the case in the Rio de Janeiro area.

select the cleanup settings¶

Once you have decided the accuracy of the OpenStreetMap data, you can setup the cleanup process:

select the accuracy of the OpenStreetMap data in the list OpenStreetMap accuracy. Values are Accurate / Not accurate
by default, GEDOT processes all the tiles that have not been cleaned yet. You can choose to process all the tiles (even those that have already be cleaned) by ticking the Process all the tiles checkbox
(optional) indicate the name of the city where the airport is located in the field Airport city (ex.: Arcachon). You can also use an OSM Geocode definition for the airport: (ex.: Santos Dumont Airport, Rio de Janeiro). This information is necessary to exclude the airport photogrammetry data, if they exist

If you select "Accurate" in the list:¶

the checkbox Keep buildings 3d data is readonly and is automatically checked
(optional) if you want to keep photogrammetry roads, tick the Keep roads 3d data checkbox (enabled by default)
(optional) if you want to keep the construction area (which can be useful if the data between the photogrammetry and OpenStreetMap are not up-to-date), tick the Keep construction area 3d data checkbox
(optional) if you want to keep residential area, without worrying about the real shapes of the buildings, tick the Keep residential area 3d data.
This is useful in the case when not all the buildings are drawn in OSM, but the residential area are clearly shaped (with the OSM tag landuse=residential and the OSM tag landuse=industrial).
It is a way of using the constructive method, even if the OSM data are not accurate. You can see an example of this type of OSM data accuracy in the Cardiff city suburbs (light grey area have the tag landuse=residential, whereas dark grey shapes represent the buildings):

If you select "Not accurate" in the list:¶

the checkbox Exclude water 3d data is readonly and is automatically checked
(optional) if you want to exclude forests 3d data, tick the Exclude forests 3d data checkbox (it roughly corresponds to the OSM tag landuse=forest)
(optional) if you want to exclude woods 3d data, tick the Exclude woods 3d data checkbox (it roughly corresponds to the OSM tag natural=wood)
(optional) if you want to exclude other ground 3d data, tick the Exclude other ground 3d data (farmlands, allotments, meadows, orchards) checkbox
(optional) if you want to exclude nature reserves 3d data, tick the Exclude nature reserves 3d data checkbox (it roughly corresponds to the OSM tag leisure=nature_reserve)
(optional) if you want to exclude parks 3d data, tick the Exclude parks 3d data checkbox (it roughly corresponds to the OSM tag leisure=park)

Run the cleanup process¶

If all the settings are correctly set, the Cleanup 3d data from Google Earth tiles button should be enabled (except if the MSFS project folder does not exist, has been renamed or removed).
Just click on this button, and a Blender window console will appear on the screen.

Once finished, you should see this on the console:

By default, the building process is executed by the MSFS fspackagetool exe. At the end of the process, you should see this window:

cleanup 3d data backup folder

The first time the cleanup process is launched on a MSFS project, a new backup folder, named "cleanup_3d_data" is created:

Arcachon
└───backup
    └───cleanup_3d_data
        └───PackageSources
            ├───arcachon-modelLib
            │   └───texture
            └───scene

This folder must be kept in order to run the process again, as it contains all the entire photogrammetry tiles (before the cleanup).
The cleanup process will always rely on those entire tiles. If the folder is deleted, and the process is executed again, it will take the already cleaned tiles, which can produce undesired results.

Cleanup results¶

Some cleanup results are presented below.

Cleanup result for a photogrammetry tile with vegetation 3d data¶

(Move the slider to display before/after comparison)

Cleanup result for a photogrammetry tile with water 3d data¶

(Move the slider to display before/after comparison)

bounding box creation

The cleaning process creates automatically a bounding box around the cleaned tile in order to preserve the dimensions of the original tile.
This ensures that the MSFS engine keeps the correct lod levels, even if the size of the tile has been reduced.

Cleanup margins

Because there can be some slight differences between the position of the OpenStreetMap buildings and roads and the corresponding photogrammetry 3d data, a margin is defined around each object to clean (moslty 8 meters). That's why it can remain some little vegetation 3d data around a road or a building.

Overall cleanup result for a photogrammetry scenery¶

In MSFS, the resulting tiles after the cleanup process are presented here (to ease the understanding, the tiles have been moved over the sea):

Fitz-Roy elevation data¶

(Move the slider to display before/after comparison)