| Home | Register | Members | Search | Links |
 |
| Thread Tools | Display Modes |
|
Guest
Posts: n/a
|
|
|
|
||
|
|
| |
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
On 21 Dic, 09:10, Rik Brown <Rik.Brown.553...@no-mx.forums.travel.com>
wrote: > claudegps;3386810 Wrote: > > > Rik, I think it's not a specific question for "GPS gurus". > > It seems more a data formats (used in mapping)... > > Do you have references explaining > > Precision Polygon Based and Centroid based that could help understand > > better your question? > > Thanks for your reply. > > To answer your question: unfortunately not. > > I have a lot of latitude/longitude pairs that for a single address are > fairly much the same, yet different. > > latitude centroid: 61.1515 > longitude centroid: -149.8227 > > latitude polygon: 61.180837 > longitude polygon: -149.816495 > > The sample street address (just happens to be a Quiznos restaurant) is: > > 3421 E Tudor Rd > Anchorage, Alaska 99507 (USA) > > I can see that the "polygon" method is more precise in a numerical sense > as it has more digits to the right of the decimal point. However, there > are discrepancies in the numbers besides the "precision" of the decimal > -- especially looking at the latitude numbers. You should figure out if the differences are related to the accuracy used to retrieve the information (which database in more precise) or if the two methods have a different reference. You could try to figure it out looking at a certain area. If the difference is (nearly) costant, then maybe they have the same information but a different reference (even if I think it's not likely... I think that everything is WGS84...) > Both ways of creating the lat/long pairs where created from the street > address above. I just can't figure out which method I should use for > mapping purposes. Does centroid and polygon data sets come from different providers? If yes, I can suppose that there are diference coming from the accuracy used to retrieve the information. You should figure out which databes is better (maybe trying checking the accuracy of several places in different areas) Anyway be prepared to have deceptive results (one DB may be better in one area and worst in another one) Im not sure I have understood what you want to do with the data anyway  Do you need the best accuracy possible, or the worst accuracy of the may satisfy your requirements? |
|
|
|
|
|||
|
Guest
Posts: n/a
|
On 12/21/2011, claudegps posted:
> On 21 Dic, 09:10, Rik Brown <Rik.Brown.553...@no-mx.forums.travel.com> > wrote: >> claudegps;3386810 Wrote: >> >>> Rik, I think it's not a specific question for "GPS gurus". >>> It seems more a data formats (used in mapping)... >>> Do you have references explaining >>> Precision Polygon Based and Centroid based that could help understand >>> better your question? >> >> Thanks for your reply. >> >> To answer your question: unfortunately not. >> >> I have a lot of latitude/longitude pairs that for a single address are >> fairly much the same, yet different. >> >> latitude centroid: 61.1515 >> longitude centroid: -149.8227 >> >> latitude polygon: 61.180837 >> longitude polygon: -149.816495 >> >> The sample street address (just happens to be a Quiznos restaurant) is: >> >> 3421 E Tudor Rd >> Anchorage, Alaska 99507 (USA) >> >> I can see that the "polygon" method is more precise in a numerical sense >> as it has more digits to the right of the decimal point. However, there >> are discrepancies in the numbers besides the "precision" of the decimal >> -- especially looking at the latitude numbers. > You should figure out if the differences are related to the accuracy > used to retrieve the information (which database in more precise) or > if the two methods have a different reference. > You could try to figure it out looking at a certain area. If the > difference is (nearly) costant, then maybe they have the same > information but a different reference (even if I think it's not > likely... I think that everything is WGS84...) >> Both ways of creating the lat/long pairs where created from the street >> address above. I just can't figure out which method I should use for >> mapping purposes. > Does centroid and polygon data sets come from different providers? > If yes, I can suppose that there are diference coming from the > accuracy used to retrieve the information. > You should figure out which databes is better (maybe trying checking > the accuracy of several places in different areas) > Anyway be prepared to have deceptive results (one DB may be better in > one area and worst in another one) > Im not sure I have understood what you want to do with the data > anyway > Do you need the best accuracy possible, or the worst accuracy of the > may satisfy your requirements? Whatever his concern is, Googling for the two terms turns up posts by Rik in two forums, so he is really concerned. BTW, Google seems to show no real information about those terms other than links to Rik's posts :-) -- Gene E. Bloch (Stumbling Bloch) |
|
|
|
|
|||
|
Guest
Posts: n/a
|
On 12/21/2011, Gene E. Bloch posted:
> On 12/21/2011, claudegps posted: >> On 21 Dic, 09:10, Rik Brown <Rik.Brown.553...@no-mx.forums.travel.com> >> wrote: >>> claudegps;3386810 Wrote: >>> >>>> Rik, I think it's not a specific question for "GPS gurus". >>>> It seems more a data formats (used in mapping)... >>>> Do you have references explaining >>>> Precision Polygon Based and Centroid based that could help understand >>>> better your question? >>> >>> Thanks for your reply. >>> >>> To answer your question: unfortunately not. >>> >>> I have a lot of latitude/longitude pairs that for a single address are >>> fairly much the same, yet different. >>> >>> latitude centroid: 61.1515 >>> longitude centroid: -149.8227 >>> >>> latitude polygon: 61.180837 >>> longitude polygon: -149.816495 >>> >>> The sample street address (just happens to be a Quiznos restaurant) is: >>> >>> 3421 E Tudor Rd >>> Anchorage, Alaska 99507 (USA) >>> >>> I can see that the "polygon" method is more precise in a numerical sense >>> as it has more digits to the right of the decimal point. However, there >>> are discrepancies in the numbers besides the "precision" of the decimal >>> -- especially looking at the latitude numbers. >> You should figure out if the differences are related to the accuracy >> used to retrieve the information (which database in more precise) or >> if the two methods have a different reference. >> You could try to figure it out looking at a certain area. If the >> difference is (nearly) costant, then maybe they have the same >> information but a different reference (even if I think it's not >> likely... I think that everything is WGS84...) >>> Both ways of creating the lat/long pairs where created from the street >>> address above. I just can't figure out which method I should use for >>> mapping purposes. >> Does centroid and polygon data sets come from different providers? >> If yes, I can suppose that there are diference coming from the >> accuracy used to retrieve the information. >> You should figure out which databes is better (maybe trying checking >> the accuracy of several places in different areas) >> Anyway be prepared to have deceptive results (one DB may be better in >> one area and worst in another one) >> Im not sure I have understood what you want to do with the data >> anyway >> Do you need the best accuracy possible, or the worst accuracy of the >> may satisfy your requirements? > Whatever his concern is, Googling for the two terms turns up posts by Rik in > two forums, so he is really concerned. > BTW, Google seems to show no real information about those terms other than > links to Rik's posts :-) OTOH, if you go here: http://en.wikipedia.org/wiki/Geocoding and search for the two terms "centroid" and "polygon" (one at a time, please!), perhaps what they have to say will offer a clue. -- Gene E. Bloch (Stumbling Bloch) |
|
|
|
|
|||
|
Guest
Posts: n/a
|
On 21 Dic, 20:53, Gene E. Bloch <blochx...@someplace.invalid> wrote:
> On 12/21/2011, claudegps posted: > > > > > > > > > > > On 21 Dic, 09:10, Rik Brown <Rik.Brown.553...@no-mx.forums.travel.com> > > wrote: > >> claudegps;3386810 Wrote: > > >>> Rik, I think it's not a specific question for "GPS gurus". > >>> It seems more a data formats (used in mapping)... > >>> Do you have references explaining > >>> Precision Polygon Based and Centroid based that could help understand > >>> better your question? > > >> Thanks for your reply. > > >> To answer your question: unfortunately not. > > >> I have a lot of latitude/longitude pairs that for a single address are > >> fairly much the same, yet different. > > >> latitude centroid: 61.1515 > >> longitude centroid: -149.8227 > > >> latitude polygon: 61.180837 > >> longitude polygon: -149.816495 > > >> The sample street address (just happens to be a Quiznos restaurant) is: > > >> 3421 E Tudor Rd > >> Anchorage, Alaska 99507 (USA) > > >> I can see that the "polygon" method is more precise in a numerical sense > >> as it has more digits to the right of the decimal point. However, there > >> are discrepancies in the numbers besides the "precision" of the decimal > >> -- especially looking at the latitude numbers. > > You should figure out if the differences are related to the accuracy > > used to retrieve the information (which database in more precise) or > > if the two methods have a different reference. > > You could try to figure it out looking at a certain area. If the > > difference is (nearly) costant, then maybe they have the same > > information but a different reference (even if I think it's not > > likely... I think that everything is WGS84...) > >> Both ways of creating the lat/long pairs where created from the street > >> address above. I just can't figure out which method I should use for > >> mapping purposes. > > Does centroid and polygon data sets come from different providers? > > If yes, I can suppose that there are diference coming from the > > accuracy used to retrieve the information. > > You should figure out which databes is better (maybe trying checking > > the accuracy of several places in different areas) > > Anyway be prepared to have deceptive results (one DB may be better in > > one area and worst in another one) > > Im not sure I have understood what you want to do with the data > > anyway > > Do you need the best accuracy possible, or the worst accuracy of the > > may satisfy your requirements? > > Whatever his concern is, Googling for the two terms turns up posts by > Rik in two forums, so he is really concerned. > > BTW, Google seems to show no real information about those terms other > than links to Rik's posts :-) That's why I was asking Rik for more details Googling I reallyhad no clues. Anyway your link on wiki has some. My interpretation is that the "polygon based" bases the location on the approximated position taking into account that a polygon (i.e. a segment of a road) includes a certain numbers of locations (i.e. from 10 to 50). So the approsimation takes the segment and puts n "10" at the start, n "50" at the end, and the other street numbers are interpolated in between. This may clear clearly lead to big errors in certain cases (10 to 50 are not spaced equally in reality) Centroid position instead should take the center of the specific parcel. This should indicate the real position of the place, but not it's position on the road. (if a parcels extends far from the road, the lat/lon indicated by teh centroid may be far from the nearest road). I think that navigation systems, are more likely to use the "polygon" based because in that way all the street numbers are easily linked to the road. (You my above interpretation at your own risk! ) Anyway if Rick has two DBs of a certian are, we could do some comparisons and maybe check if my interpretation is correct. |
|
|
|
|
|||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
|
| |
|
| Thread Tools | |
| Display Modes | |
|
|
Linear Mode
