• Which the release of FS2020 we see an explosition of activity on the forun and of course we are very happy to see this. But having all questions about FS2020 in one forum becomes a bit messy. So therefore we would like to ask you all to use the following guidelines when posting your questions:

    • Tag FS2020 specific questions with the MSFS2020 tag.
    • Questions about making 3D assets can be posted in the 3D asset design forum. Either post them in the subforum of the modelling tool you use or in the general forum if they are general.
    • Questions about aircraft design can be posted in the Aircraft design forum
    • Questions about airport design can be posted in the FS2020 airport design forum. Once airport development tools have been updated for FS2020 you can post tool speciifc questions in the subforums of those tools as well of course.
    • Questions about terrain design can be posted in the FS2020 terrain design forum.
    • Questions about SimConnect can be posted in the SimConnect forum.

    Any other question that is not specific to an aspect of development or tool can be posted in the General chat forum.

    By following these guidelines we make sure that the forums remain easy to read for everybody and also that the right people can find your post to answer it.

DETECTFEATURES for Buildings

B

BrBlazer

Messages
481
Hey all,

We've all seen the awesomeness that is the detectfeatures for the vegetation, but I would like to know if it's currently or can be in the future, possible to detect buildings from the "rooftops" on images.

I know trees are mostly standard green which makes easier detection as building rooftops have various colors, but if we can select one or more color ranges that would determine that those colors belong to rooftops. I guess the grey rooftops would be the most difficult, since it can be "confused" for sidewalks, roads, car parks, etc... so we would need to limit the size of what could be considered a building.

It would be a great way as I am still trying to figure this out: http://www.fsdeveloper.com/forum/th...g-footprints-from-imagery.428980/#post-664745

Kindly,

Blazer
 
Hi,

I was contacted by a developer a few weeks ago who used it successfully to extract building footprints as well. Only the autogen objects didn't turn out correctly for the heading. That's related to the way minimum area polygon is calulated. Detected features only have 90 degree angles (due to the I pixels). To fix that I have a feature request on the wishlist to optimize polygons.
 
That is wonderful news. I have been trying this since the post I made and a got an ok looking filter for my test location, on a small exported TIF file 2031x1995 it goes fine, however, if I insert the whole image, 6827x4096 I end up with this error handler report:
An unexpected error occured during running your configuration

Error:
Attempt to create 0x1024 dataset is illegal,sizes must be larger than zero.


EDIT: I used global mapper to split the main file into 2048x2048 file and then they executed fine. After all further processing to get the final AGN file, the result was very promising, apart from the buildings being the same heading as you mentioned already.
 
Last edited:
You have a SPLITGRID command in your configuration? With which size?

It seems somehow the code tries to make a 0 pixels width bitmap and GDAL doesn't like that.
 
I tried with both SPLITGRID|1|* and SPLITGRID|AGN|*

The "1" takes longer to process but ultimately both have this error.
 
Hi,

OK, I guess you are just unlucky that a very small part of your image fits into another cell and that that part is 0 pixels in size. I'll try to reproduce it here and then add a check to prevent the crash.
 
Hello :-)

It's been a while since I worked with this for extracting autogen. Is there any new developments in relation to detect and extract buildings from images to make autogen?
 
Hi,

Not that much has changed yet. The features that are detected are simplified now, so you should get a better building orientation.

Another idea recently mentioned is to do the detection on a digital map that shows the footprints, instead of from imagery. That might give better results, but I haven't tested it yet.

I would say, just try and if you have trouble I'll be happy to help.
 
You must log in or register to reply here.
Back
Top