Re: [Talk-ca] Importing buildings in Canada

2020-03-22 Thread Daniel @jfd553
Lol

Sent from Galaxy S7


From: Tim Elrick 
Sent: Sunday, March 22, 2020 6:49:48 PM
To: Daniel @jfd553 
Cc: talk-ca@openstreetmap.org 
Subject: Re: [Talk-ca] Importing buildings in Canada

Hi Daniel,

I agree with you. I didn't pay attention to the fact that Squamish is
located in a hilly area.

Greetings from Quebec's flatlands,
Tim

On 2020-03-22 14:16, Daniel @jfd553 wrote:
Hi all, sorry for this long Email.

Thanks to Tim to have comment! He wrote: “I [...] found that you can
either align the hospital with the underlying imagery or the houses to
the right of the task, but not both at the same time. [...]  If we
assume that the aerial imagery data is the correctly projected [...], we
would have to correct the position of all the buildings according to the
underlying aerial imagery.”

Well, you are right. Actually, I did not align most of the buildings to
the image! Why? Because unless proven otherwise, ODB data should be more
accurate (XY) than most images available, especially in hilly areas.
Municipalities generally use aerial photos to create their maps (ODB
data). Because these aerial photos provide multiple views of the same
area, they can be used to compute digital elevation models (DEMs)
showing even buildings’ height. Only once done, they can create accurate
ortho-images (orthographic view [1]). Without an accurate DEM, objects
location on an image is not accurate either, because we are in a
perspective view [1].
The DEMs used to create available OSM images generally do not have a
sufficient accuracy in mountainous areas. This is the case of the
Squamish area where the image shows many examples of perspective views
[1]. In flat areas, this effect is minimal, which makes it possible to
adjust an image over a large region with a great accuracy. The only
visible effect is then related to buildings’ height.

Regarding the hospital, it is located on a hill between two plateaus.
The image can be adjusted with a good accuracy on the flat area near the
river, or on the plateau on the top of the hill (potentially with
another offset), but it is more difficult in between. I tried to adjust
its geometry (details) from its original ODB location.

I adjust the image to surrounding buildings when I need to map a new one
or add details to an existing one. I may also look at available GPS
tracks to confirm general ODB data location.

Thanks again. Comments?
Daniel

[1] https://en.wikipedia.org/wiki/Orthophoto

___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca


Re: [Talk-ca] Importing buildings in Canada

2020-03-22 Thread Tim Elrick

Hi Daniel,

I agree with you. I didn't pay attention to the fact that Squamish is 
located in a hilly area.


Greetings from Quebec's flatlands,
Tim

On 2020-03-22 14:16, Daniel @jfd553 wrote:
Hi all, sorry for this long Email.

Thanks to Tim to have comment! He wrote: “I [...] found that you can 
either align the hospital with the underlying imagery or the houses to 
the right of the task, but not both at the same time. [...]  If we 
assume that the aerial imagery data is the correctly projected [...], we 
would have to correct the position of all the buildings according to the 
underlying aerial imagery.”


Well, you are right. Actually, I did not align most of the buildings to 
the image! Why? Because unless proven otherwise, ODB data should be more 
accurate (XY) than most images available, especially in hilly areas.
Municipalities generally use aerial photos to create their maps (ODB 
data). Because these aerial photos provide multiple views of the same 
area, they can be used to compute digital elevation models (DEMs) 
showing even buildings’ height. Only once done, they can create accurate 
ortho-images (orthographic view [1]). Without an accurate DEM, objects 
location on an image is not accurate either, because we are in a 
perspective view [1].
The DEMs used to create available OSM images generally do not have a 
sufficient accuracy in mountainous areas. This is the case of the 
Squamish area where the image shows many examples of perspective views 
[1]. In flat areas, this effect is minimal, which makes it possible to 
adjust an image over a large region with a great accuracy. The only 
visible effect is then related to buildings’ height.


Regarding the hospital, it is located on a hill between two plateaus. 
The image can be adjusted with a good accuracy on the flat area near the 
river, or on the plateau on the top of the hill (potentially with 
another offset), but it is more difficult in between. I tried to adjust 
its geometry (details) from its original ODB location.


I adjust the image to surrounding buildings when I need to map a new one 
or add details to an existing one. I may also look at available GPS 
tracks to confirm general ODB data location.


Thanks again. Comments?
Daniel

[1] https://en.wikipedia.org/wiki/Orthophoto


___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca


Re: [Talk-ca] Importing buildings in Canada

2020-03-22 Thread Daniel @jfd553
Hi all, sorry for this long Email.

Thanks to Tim to have comment! He wrote: “I [...] found that you can either 
align the hospital with the underlying imagery or the houses to the right of 
the task, but not both at the same time. [...]  If we assume that the aerial 
imagery data is the correctly projected [...], we would have to correct the 
position of all the buildings according to the underlying aerial imagery.”

Well, you are right. Actually, I did not align most of the buildings to the 
image! Why? Because unless proven otherwise, ODB data should be more accurate 
(XY) than most images available, especially in hilly areas. 
Municipalities generally use aerial photos to create their maps (ODB data). 
Because these aerial photos provide multiple views of the same area, they can 
be used to compute digital elevation models (DEMs) showing even buildings’ 
height. Only once done, they can create accurate ortho-images (orthographic 
view [1]). Without an accurate DEM, objects location on an image is not 
accurate either, because we are in a perspective view [1]. 
The DEMs used to create available OSM images generally do not have a sufficient 
accuracy in mountainous areas. This is the case of the Squamish area where the 
image shows many examples of perspective views [1]. In flat areas, this effect 
is minimal, which makes it possible to adjust an image over a large region with 
a great accuracy. The only visible effect is then related to buildings’ height. 

Regarding the hospital, it is located on a hill between two plateaus. The image 
can be adjusted with a good accuracy on the flat area near the river, or on the 
plateau on the top of the hill (potentially with another offset), but it is 
more difficult in between. I tried to adjust its geometry (details) from its 
original ODB location. 

I adjust the image to surrounding buildings when I need to map a new one or add 
details to an existing one. I may also look at available GPS tracks to confirm 
general ODB data location.

Thanks again. Comments?
Daniel

[1] https://en.wikipedia.org/wiki/Orthophoto
___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca


[Talk-ca] weeklyOSM #504 2020-03-10-2020-03-16

2020-03-22 Thread weeklyteam
The weekly round-up of OSM news, issue # 504,
is now available online in English, giving as always a summary of a lot of 
things happening in the openstreetmap world:

 https://www.weeklyosm.eu/en/archives/12980/

Enjoy! 

Did you know that you can also submit messages for the weeklyOSM? Just log in 
to https://osmbc.openstreetmap.de/login with your OSM account. Read more about 
how to write a post here: 
http://www.weeklyosm.eu/this-news-should-be-in-weeklyosm 

weeklyOSM? 
who: https://wiki.openstreetmap.org/wiki/WeeklyOSM#Available_Languages 
where?: 
https://umap.openstreetmap.fr/en/map/weeklyosm-is-currently-produced-in_56718#2/8.6/108.3
___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca