[Talk-ca] Batiments - Fonctions orthogonales

Pierre Béland pierzenh at yahoo.fr
Sat Jun 1 03:13:38 UTC 2019


Bonjour Jarek
extring === external ring groupe de batiments "building.geojson",  batiments individuels
exemple  https://www.openstreetmap.org/way/661895520
"building_extring.geojson" batiments adjacents qui partagent des nodesexemple http://overpass-turbo.eu/s/JAe
geojson si on clique sur polygone
vars :
| ways_id | 661895520,639843855,661895517,661895518,661895519,661895521,661895522,661894733 |
| grp_poly | 7 |


"building_extring_orthogonal.geojson"
Orthogonal === Avec correction orthogonale (squarred result)  A noter version préliminaire et améliorations a venircorrection orthogonale - actuellement tous les batiments - correction (squarring) si 80 < angle < 100 - aucun filtre actuellement 
building_orthogonal.geojson  correction batiments individuels  - on observe parfois rotation du batiment - cela pourrait être corrigé en identifiant aligmement général du batiment (coté le plus long ?)   et aligner ensuite autres cotés.  JOSM, code java fonction orthogonale, si je comprends que c'est la procédure suivie


|  * Estimate the direction of the segments, given the first segment points in the |
|


|  * direction <code>pInitialDirection</code>. |
|


|  * Then sum up all horizontal / vertical segments to have a good guess for the |
|

 * heading of the entire way.
https://github.com/stefanocudini/orthogonalize-js/blob/master/OrthogonalizeAction.josm.java

 
Pierre 
 

    Le vendredi 31 mai 2019 17 h 49 min 21 s UTC−4, Jarek Piórkowski <jarek at piorkowski.ca> a écrit :  
 
 Hi Pierre,

Thanks for sending these out.

Can you briefly confirm what the "building.geojson",
"building_extring.geojson", "building_extring_orthogonal.geojson"
files represent? I'm not really familiar with the terms, perhaps
because I don't have much of a GIS or geometry background. Is the
"extring" file only those buildings that don't have superfluous nodes?
"extring_orthogonal" contains only those that are square and don't
have superfluous nodes?

I guess OSM data is used for easy testing? I remain very interested as
to how the Statcan building footprints for that area look like when
cleaned up - I hope for better accuracy than trying to estimate from
low-res or off-vertical imagery.

It doesn't help that Github GeoJSON preview evidently uses a super-old
version of OSM data for base map...

Thanks again,
--Jarek  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-ca/attachments/20190601/9479bb0c/attachment.html>


More information about the Talk-ca mailing list