[talk-ph] Fwd: RFC: importing Palawan roads

maning sambale emmanuel.sambale at gmail.com
Mon Jun 8 03:48:10 BST 2009


Excellent!  May GPS ba sya?  Can he look at the donated Palawan data?

On Mon, Jun 8, 2009 at 10:38 AM, George Tujan<gtujan at gmail.com> wrote:
> @maning: i've contacted a former officemate who lives in palawan and
> is every interested in joining and contributing at least may palawan
> mapper na tayo. :). His name is Jefferson Fermo and he was very
> excited when he saw the current state of OSM in the Philippines :)
>
>
> On Sat, Jun 6, 2009 at 3:13 PM, maning
> sambale<emmanuel.sambale at gmail.com> wrote:
>> With more eyeballs all bugs are seen (Linus Torvalds)
>>
>>> Doing a quick check on proportions and shapes of different tracks from different portion of
>>> the map, I think the data supplied is trash. :-(
>>
>> As per Rally's assessment, I am holding-off to importing this data.
>> Unless, others can make sense of how we can use this for OSM.  I can
>> provide a download link to anyone who wants to play around with it.
>> If you think it's useable for OSM, we can discuss this here.
>>
>> Personally, nanghinayang lang ako, if this is good data (or at least
>> useful) a good chunk of roads will be added to OSM.  But as Rally
>> said,
>>> So, we'll surely get in trouble or lose good reputation for uploading bad data, hehe
>>
>> I'd rather accept someone saying we have wrong data because of an
>> honest newbie experimenting around than importing data we know having
>> big problems.
>>
>> Thanks Rally for doing the assessment.  Any other comments?
>>
>>
>> On Sat, Jun 6, 2009 at 11:49 AM, Eugene Alvin Villar<seav80 at gmail.com> wrote:
>>> I was about to ask you to ask them how they compiled their data. It
>>> obviously wasn't traced over Google Earth nor taken using GPS tracks. I
>>> think Rally might have guessed their method correctly.
>>>
>>>
>>> On Sat, Jun 6, 2009 at 11:26 AM, maning sambale <emmanuel.sambale at gmail.com>
>>> wrote:
>>>>
>>>> ---------- Forwarded message ----------
>>>> From: Rally de Leon <ralleon at gmail.com>
>>>> Date: Fri, Jun 5, 2009 at 9:36 PM
>>>> Subject: Re: [talk-ph] RFC: importing Palawan roads
>>>> To: maning sambale <emmanuel.sambale at gmail.com>
>>>>
>>>>
>>>> Doing a quick check on proportions and shapes of different tracks from
>>>> different portion of the map, I think the data supplied is trash. :-(
>>>>
>>>> I've seen shape files very much like this (like the one I obtained
>>>> from Taytay Assessor) that was just traced over TAX MAPPING paper
>>>> sheets (purely based on shape guestimation and visual
>>>> ratio-and-proportion) then converted then to autocad drawings. I've
>>>> seen them how they draw the lot shapes on paper (during the time
>>>> nagtatambay ako sa assessor). It's doable if you know where to chop
>>>> the areas (eg. per subdivision or per purok), then convert it back to
>>>> raster image, then georectify that image (for every area) which will
>>>> need actual gps waypoints, then trace your vector lines over the
>>>> raster layer. But what's the point? we are talking of the whole
>>>> palawan.The purpose of this data is just for assessment and for
>>>> assigning PIN codes on the tax declaration, not for navigation nor for
>>>> making proportionally correct paper map.
>>>>
>>>> I highly doubt that the data came from actual field survey. Because
>>>> I've traced Namria topo maps before (in Luzon datum), and they were
>>>> proportionally correct when I translated them to WGS84. My old
>>>> personal garmin maps were made this way with fairly high degree in
>>>> accuracy. But in this case, the shp file have wrong proportions,
>>>> different shapes, different angles, a lot of extraneous lines and
>>>> streets (that doesn't make sense), really bad. I tried Luzon-Mean,
>>>> Luzon-Philippines, Luzon-Mindanao (all failed), so I tried to cheat by
>>>> just dragging them over google satellite image, still failed, won't
>>>> fit. No wonder you can't make it fit using all available datum, hehe.
>>>> nadaya yata tayo dun. If we are lucky enough to have one subdivision
>>>> fit the roads of a satellite image (yahoo or google), the adjacent
>>>> subdivision won't (as i've tried it since this was really easy to do
>>>> in trackmaker pro). In my experience, even those subdivisions as far
>>>> as 500meters from your reference tracklog should still fit the
>>>> satellite image with your extended vector traces; in this case, even
>>>> the neighboring streets a few blocks away doesn't.
>>>>
>>>> So, we'll surely get in trouble or lose good reputation for uploading
>>>> bad data, hehe. I suggest we just recruit gps-tracker volunteers in
>>>> any of the streets in palawan (where there are good hi-res satellite
>>>> images), and lets do it the old fashion way.
>>>>
>>>> Call back the source of this data and ask him to confess (if they
>>>> really made that survey). kung ayaw umamin, i-torture natin ;-D
>>>>
>>>>
>>>> On Fri, Jun 5, 2009 at 6:17 PM, maning sambale
>>>> <emmanuel.sambale at gmail.com> wrote:
>>>> >
>>>> > I tried every projection parameters I can think of pero may shift pa
>>>> > rin, anyway the shift is systematic so moving the whole data to fit
>>>> > existing OSM roads, should be straightforward.
>>>> >
>>>> > Here's the data, around 7 MB:
>>>> > http://esambale.wikispaces.com/file/view/palawan_roads.zip
>>>> >
>>>> > Tirahin mo sya!
>>>> >
>>>> > On Fri, Jun 5, 2009 at 6:01 PM, Rally de Leon<ralleon at gmail.com> wrote:
>>>> > > can you send me a copy of this palawan tracks? how big? i'm sure you
>>>> > > already
>>>> > > have your methods of converting to wgs84.
>>>> > >
>>>> > > but let me see what i can do (para meron naman akong mapaglibangan
>>>> > > this
>>>> > > weekend). don't worry, i'm not going to trace over googleearth. (lets
>>>> > > just
>>>> > > compare notes later) :-)
>>>> > >
>>>> > > i'm just going to try if it's as easy as shifting the entire thing on
>>>> > > the
>>>> > > built-in datum-shifter of GPS trackmaker, if it works and passed our
>>>> > > counterchecks,
>>>> > > then the data goes to JOSM for additional validations, edits and
>>>> > > tagging.
>>>> > >
>>>> > > else if it's not that easy, maybe we can use a method of shifting
>>>> > > arbitrary
>>>> > > control points. (most likely, palawan is on Zone 2 of Luzon Datum,
>>>> > > then
>>>> > > convert to WGS84) (practically, all main Luzon island is on Zone 3 -
>>>> > > the one
>>>> > > used by surveyors). The alignment of these "luzon-datum tracks" to the
>>>> > > "computed WGS84 control points" can be done entirely within JOSM by
>>>> > > just
>>>> > > dragging the entire tracks. We have many ways to check the accuracy
>>>> > > afterwards. medyo barok (as we will use graphical method - JOSM's is
>>>> > > very
>>>> > > good at this), but it may work without the need to for mathematical
>>>> > > conversions and without the need to rely on googleearth alignments.
>>>> > >
>>>> > > On Fri, Jun 5, 2009 at 2:51 PM, maning sambale
>>>> > > <emmanuel.sambale at gmail.com>
>>>> > > wrote:
>>>> > >>
>>>> > >> > There is no online map we can compare the traces to to see how
>>>> > >> > accurate
>>>> > >> I'm doing it when I got this message. :)
>>>> > >>
>>>> > >> See for yourself over GEarth (blue is OSM, red is the road data).
>>>> > >> http://farm4.static.flickr.com/3566/3596753363_51c00bb5f3_o.jpg
>>>> > >>
>>>> > >> Notice the shift?
>>>> > >>
>>>> > >> (copyright image of DigitalGlobe and Google included to abide by it's
>>>> > >> Terms of Use) :)
>>>> > >>
>>>> > >>
>>>> > >> On Fri, Jun 5, 2009 at 2:18 PM, Andre
>>>> > >> Marcelo-Tanner<andre at enthropia.com>
>>>> > >> wrote:
>>>> > >> > There is no online map we can compare the traces to to see how
>>>> > >> > accurate
>>>> > >> > it is?
>>>> > >> >
>>>> > >> > _______________________________________________
>>>> > >> > talk-ph mailing list
>>>> > >> > talk-ph at openstreetmap.org
>>>> > >> > http://lists.openstreetmap.org/listinfo/talk-ph
>>>> > >> >
>>>> > >>
>>>> > >>
>>>> > >>
>>>> > >> --
>>>> > >> cheers,
>>>> > >> maning
>>>> > >> ------------------------------------------------------
>>>> > >> "Freedom is still the most radical idea of all" -N.Branden
>>>> > >> wiki: http://esambale.wikispaces.com/
>>>> > >> blog: http://epsg4253.wordpress.com/
>>>> > >> ------------------------------------------------------
>>>> > >>
>>>> > >> _______________________________________________
>>>> > >> talk-ph mailing list
>>>> > >> talk-ph at openstreetmap.org
>>>> > >> http://lists.openstreetmap.org/listinfo/talk-ph
>>>> > >
>>>> > >
>>>> >
>>>> >
>>>> >
>>>> > --
>>>> > cheers,
>>>> > maning
>>>> > ------------------------------------------------------
>>>> > "Freedom is still the most radical idea of all" -N.Branden
>>>> > wiki: http://esambale.wikispaces.com/
>>>> > blog: http://epsg4253.wordpress.com/
>>>> > ------------------------------------------------------
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> cheers,
>>>> maning
>>>> ------------------------------------------------------
>>>> "Freedom is still the most radical idea of all" -N.Branden
>>>> wiki: http://esambale.wikispaces.com/
>>>> blog: http://epsg4253.wordpress.com/
>>>> ------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> talk-ph mailing list
>>>> talk-ph at openstreetmap.org
>>>> http://lists.openstreetmap.org/listinfo/talk-ph
>>>
>>>
>>>
>>> --
>>> http://vaes9.codedgraphic.com
>>>
>>
>>
>>
>> --
>> cheers,
>> maning
>> ------------------------------------------------------
>> "Freedom is still the most radical idea of all" -N.Branden
>> wiki: http://esambale.wikispaces.com/
>> blog: http://epsg4253.wordpress.com/
>> ------------------------------------------------------
>>
>> _______________________________________________
>> talk-ph mailing list
>> talk-ph at openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/talk-ph
>>
>



-- 
cheers,
maning
------------------------------------------------------
"Freedom is still the most radical idea of all" -N.Branden
wiki: http://esambale.wikispaces.com/
blog: http://epsg4253.wordpress.com/
------------------------------------------------------




More information about the talk-ph mailing list