[Talk-bd] The Kamal Ataturk Dual Carriageway

Syed Ishtiaque Ahmed ishtiaque at csebuet.org
Fri Jan 27 19:10:42 GMT 2012


Thank you Larry for your contribution. I am sure this will help our young
mappers.

ishtiaque

On Tue, Jan 24, 2012 at 6:12 AM, Larry O'Neill <larryoneill at gmail.com>wrote:

> AAAH, the screenshots are too big for the email server to allow them.
> I'll try to find a handy online location for them later so that you can
> see them.
> Or, you could open JOSM and look at the area around Banani Market and
> Gulshan2 Circle....
>
> Larry
>
>
> On Tue, Jan 24, 2012 at 11:09 AM, Larry O'Neill <larryoneill at gmail.com>wrote:
>
>> Hi Guys,
>>
>> This is my first descriptive document on my edits in Bangladesh, and I
>> certainly hope it wont be the last.
>> On the 7th of January, I submitted  osm.org/browse/changeset/10322877
>> This is my description of why and how this was done.
>>
>> Dhaka has alot of dual carriageways - more than many cities I have been
>> in.
>> Navigating through Dhaka will almost always involve using a dual
>> carriageway road at some point.
>> This can introduce some interesting navigational challenges, as the
>> physical barrier between the two carriageways imposes obvious restrictions.
>> In many cases in Dhaka, turning right at a junction with a DC is not an
>> option, left is the only way you can go.
>> This reality is not really reflected if the dual carriageway is mapped as
>> a single carriageway.
>> An example of this exists here:
>> osm.org/?lat=23.76438&lon=90.37224&zoom=16
>> Mirpur road is marked on the map as a single carriageway, and so it is
>> reasonable to assume that coming from Aurangajeb Road, you can drive
>> straight to Lake Road, or turn right to go south on Mirpur Road.
>> The concrete wall would decide otherwise if you tried this, and so the
>> map doesnt really reflect what's on the ground.
>> Navigational software based on this data will trust the data, and so the
>> data needs to reflect the reality.
>>
>> This was the case with Kamal Ataturk, and so it was changed.
>> The editing of it was relatively simple, but it did rely on the use of
>> the Bing aerial imagery, and may not be entirely correct.
>> The Bing imagery of Dhaka is an incredibly valuable resource
>> [nb: I am a JOSM user, I havent dont any of this in potlatch, any
>> potlatch users please feel free to reply with equivalent descriptions...]
>> 1) open JOSM
>> 2) go to File menu, Download from OSM. OR hit the download button on the
>> main toolbar. OR hit Ctrl+Shift-D
>> 3) select the area containing the road to be edited
>> 4) download and allow it to render the data.
>> 5) go to Imagery menu, Bing Sat.
>> 6) after the background image has loaded, it may need to be moved
>> slightly. Imagery menu, New offset.
>> 7) use this to realign the imagery as well as you can. You may need to
>> match it against existing OSM data or gps traces. Click and drag on the
>> main editing pane.
>> 8) hopefully the barrier between the two sides of the DC road will be
>> visible, click and drag the current road to one side of the barrier for its
>> full length. Realign or add nodes if necessary.
>> 9) after moving the existing data to one side for the full length of the
>> road, click the draw nodes button on the edit toolbar, and draw in the
>> opposite carriageway for the full length.
>> 10) make sure the direction of the existing way is correct, and change if
>> necessary [click on a way and hit r to reverse it]. Add the oneway=yes tag
>> to all sections of the existing way.
>> 11) select the existing way and hit Ctrl+c to copy it. Click on your new
>> way and hit Ctrl+Shift-V to paste the tags. This saves a lot of time.
>> 12) edit any junctions along the road, showing where a minor road joins
>> both carriageways, and where it only joins one side.
>> 13) make sure that all possible movements at either and of the DC are
>> maintained
>> 14) upload the data and you're done!
>>
>> My first screenshot above shows the two junctions either side of Banani
>> Market, where cars can be seen turning through a gap in the wall.
>> The junction west of this reflects the restriction the continuous wall
>> imposes.
>> My second screenshot shows how the Gulshan end of the DC was terminated,
>> allowing for all possible movements.
>> This could have been achieved by joining all of the nodes at a single
>> point in the centre of Gulshan circle.
>> While that would have been topologically correct, and some would argue in
>> favor of mapping it that way, it is my personal preference to do it as
>> shown.
>> Note also in my second screenshot the poorly aligned Bing imagery. This
>> kind of alignment should be corrected before using it to edit.
>> Another thing to note is the fact that the circular road around the
>> Gulshan junction is cut a the two dual carriageway sections on  the south
>> and west.
>> The northern and eastern roads leading out of this junction are still
>> incorrectly mapped as single carriageways.
>> After doing this once, you may find it surprisingly quick to do it for
>> another road in Dhaka or any other Bangladeshi town or city.
>>
>> If any of my edits on the Kamal Ataturk are incorrect, have restrictions
>> where none exist, or would allow movement that cant be made, please feel
>> free to change it.
>> This is the beauty of collaborative and crowd sourced data - if it's
>> wrong, correct it and move on! =0)
>>
>> I hope that the above description will be of some help to those that are
>> new to the OSM project, and that we will soon see more Dual Carriageways
>> mapped in Dhaka!
>> Any comments, questions, criticisms, improvements or otherwise are
>> welcome.
>>
>> Thanks
>> Larry
>>
>>
>
> _______________________________________________
> Talk-bd mailing list
> Talk-bd at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-bd
>
>


-- 
Syed Ishtiaque Ahmed
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-bd/attachments/20120127/bc0fd028/attachment.html>


More information about the Talk-bd mailing list