• 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.

P3D v5 Combine TFE's tf2 files in one large database?

Clutch Cargo

Clutch Cargo

Messages
1,606
Country
unitedstates
Arno,
I have dozens of tf2 files based on hundreds of square kilometers, thousands a image samples and thousands of sample points. As a test, I would like to combine them into one large tf2 file - sort of like the large database we were talking about and see if this provides enough data to detect properly in the geographic area I am working on (currently the state of California, USA). I imagine a little app or selection in scenProc to combining tf2's into one would be the best way. And I assume creating this large of a file may take hours to process? One that would need to be ran on its own and not part of a script with lots of other steps (like combining and then immediately into detection and exports). If it shows promising results, this could be an option for the end-user to create their own data bases rather than waiting on you to collect 10's of 1,000's of samples.

Question - one observation I have noticed, when I continue adding sample images and sample points, as I click on "save" it will take a little longer each time to run the algorithm. It starts out instantly saving and then a few seconds longer as I continue to add. My typical wait time can run up to 1-2 minutes each time I save/update a file.... that's typically a few dozen images and 100's of sample points. That's fine, I am used to that and not sure if there is a way to improve the processing speed. I again suspect the process of combining dozens of files which may take several hours to process (hopefully no crashes!). My question is once combined into a single tf2, the processing of a detection script will also take hours for detection or all the hard work has been done beforehand? Or, running the tf2 detection time-wise will be somewhat "normal"? For me "normal" is running detection at about 7 minutes per a 30 square kilometer image. Which is ok but if it takes several hours per image... well this process would not be applicable. I guess the test will reveal.

This test is for vegetation detection only. If successful, then try the same for water detection.

Thoughts on this idea?
 
Hi,

Are all the TF2 files the same except for the sample images and sample points? Or are the steps also different.

If you combine multiple files that are trained on a specific subset you might see that the combination gives different detection results. Unless the different areas are similar, but in that case you would not have to make a separate filter for it.

If you use the SVM step the learning of the algorithm is done in the editor and not at runtime in the script. But for other steps it is done at runtime. So hard to say what it will do for the processing time without knowing the details of the filter.
 
I would have to check to ensure all are the same. I know at the beginning I was trying lots of different steps but I settled with this one I have posted. It tended to give me overall good results with the least processing time. I see I did have included the SVM step. What I could always do is choose a test area, combine 2-3 tf2's and check results... then add a few more and check results, etc.
Current Veg Detection Steps.jpg
 
You must log in or register to reply here.
Back
Top