search

openAIP / openaip

Public issue tracker of www.openaip.net.
39 stars 3 forks source link

Update german airspace, it is old. #319

Closed bubeck closed 5 months ago

bubeck commented 9 months ago

There are updates for german airspace. Please import again. I also provide a GitHub repo at https://github.com/bubeck/airspace_germany that always offers the latest version. If you like, you can pull in from there.

reskume commented 9 months ago

Hi,

can you please be more precise in what is outdated? The latest incremental update to the German airspace was done on 14.11.2023 and added 4 new temporary airspaces. Since updates to the German airspace are also mostly done by Peter (BIG kudos to him) there shouldn't be any difference to the officially available file of the DAeC. Actually the openaip data should be more precise since we add also temporary airspaces that normally are not found in the DAeC file.

bubeck commented 9 months ago

I can only give you an example at https://www.openaip.net/map#15.38/48.779928/8.883274 Here you see, that the airspaces are crossing each other.

I sent fixes to Peter Kemme (is this the same Peter you mention here?) and he incorated them ninto his openair file of germany as published at DAeC. But still, the problem is visible in openaip.net. If you look at airspaces.xcontest.org, you see, that they are fixed there (because I updated the latest openair file there).

Is there a way to download (german) openair file from openaip.net, so that I can take a look at the underlying defintion?

reskume commented 9 months ago

You can log into openaip. When logged in, you can filter the airspace list to your needs, i.e. set country=DE. Then you can export into either standard or extended openair format.

bubeck commented 9 months ago

one problem is in

AC D AN Stuttgart 128.950 AH FL100 AL 4500ft MSL

openaip.net has DP 48:38:54 N 009:37:28 E

But the newest airspace_germany.txt has DP 48:38:54 N 009:37:25 E

So, it seems, that openaip is not using the latest version.

This is an indication, that you do not use the latest version as published on DAEc (2023-10-05).

bubeck commented 7 months ago

Any update on this?

reskume commented 7 months ago

I reviewed the dataset and also talked to Peter about this issue. From what I understand, the problem is not an invalid or outdated airspace but the fact that airspace geometries overlap at certain points. If this is the case, I would like to close this issue since this "problem" cannot be solved with our few resources. Most of the source data is in openair which defines arcs, circles etc and everything is converted into geometries (polygons) when imported into the openAIP database. Those polygons, when created, will consist of a finite number of points and especially arcs/circles will only be approximations. Due to that, the openAIP dataset will always have overlapping airspaces.

There are GIS tools that can fix the topography of a dataset but when done, it will definitely change the complete dataset accordingly. QA is not given in this case and we don't have the manpower to create and review a complete country dataset from scratch which would takes us an enormous amount of time.

bubeck commented 7 months ago

Thanks for your answer. You are right and wrong.. :-)

  1. This issue is about outdated data in openaip only. It seems, that you simply are not using the 2023-10-05 version. To "prove" this, I selected one changed point (see above) to verify, that you are using a previous data set. This is the content of this issue.
  2. Yes, there are problems in the dataset itself. I wrote a parser, verifiyer and visualizer to help fixing these problems (https://github.com/bubeck/airspace_germany/blob/main/bin/check-consistency.py) and I sent fixes to Peter, whenever I found a fix. Current data problems, are sometimes overlapping polygons, but this is due to an error of swapped points. If you do an error and swap two adjacent points, then you get a twist. This is very very small, and no one has probably ever seen it. Other types of errors are also there. They are all very very small and not big. I am waiting for a response from him for my latest input, so I assume he mixes "1" and "2". However, I dont want you to fix these problems in openaip, as they must be fixed upstreams. I work on this together with Peter to improve data upstreams.
  3. To help you (and others) to prevent "1" (outdated data), I created https://github.com/bubeck/airspace_germany to have easy access to Peters Airspace defintion and see what changes he did.

I am willing to improve data quality in the first place, so that we all have the best data possible. Thanks for your help with openaip.

reskume commented 5 months ago

Airspace for Germany received the new 03/2024 update today.

bubeck commented 5 months ago

Thanks!