[mapserver-users] performance problem with ortho images

[Rahkonen Jukka]

Hi,
 
Oh yes, my images are always covering full mapsheets which are rectangles.  They are naturally polygons, even simple sort of, but in your case you should use more accurate methods for finding out which images from the 2 m tileindex file could be removed.  Idea remains the same, if you know beforehand that some images will be left under another image layer you can remove them from the tileindex so they are not wasting Mapserver processing time.  And the goal is to have only 2 layers:
2m data shows between scale 200k to 100k
2 m data + 20 cm data combined and unnecessary 2 m images removed show between 100k and 500.
 
Or then you can use just the second layer as "best available image".
 
Same tileindex file can contain images of both resolution, and in places where images are overlapping the one which has bigger fid (feature id) in the shapefile will be drawn last. 
 
-Jukka-

 

________________________________

	Lähettäjä: Liang Jian [mailto:Liang.Jian at eniro.com] 
	Lähetetty: 16. marraskuuta 2009 15:40
	Vastaanottaja: Rahkonen Jukka; mapserver-users at lists.osgeo.org
	Aihe: RE: [mapserver-users] performance problem with ortho images
	
	
	Hi Jukka,
	 
	I'm not sure that I understand you correctly. I tried to clip 2m data coverage by 20cm data- get some holes in 2m data index file. And mapfile setup as follow: 
	 2m data shows between scale 200k to 100k
	 2m data with holes show between 100k to 500
	 20cm data shows between 100k to 500.
	 
	It doesn't help. I don't think mapserver can read polygons for tileindex. It sounds like using rectangle.
	 
	/Liang

________________________________

	From: Rahkonen Jukka [mailto:Jukka.Rahkonen at mmmtike.fi] 
	Sent: den 16 november 2009 13:36
	To: Liang Jian; mapserver-users at lists.osgeo.org
	Subject: Re: [mapserver-users] performance problem with ortho images
	
	
	Hi,
	 
	Is it necessary to show 2 m data at all in the areas where 20-50 cm imagery is available?  If not you can first make tileindex shapefiles from 2 m images and 20-50 cm images.  Open both shapefiles with for example OpenJUMP and select and remove those polygons from 2 m shapefiles which are totally covered by you 20-50 cm tiles.  Combine the two layer, save and use the result as a tileindex file for Mapserver.
	 
	There may be problems in the areas with overlapping images. To get the desired image to show on top it may be necessary to rewrite the shapefile by so that it will be ordered so that the small pixel size images will be written last in the shapefile.  Then they should be drawn on top of 2 m imagery.
	 
	If you want the show 2 m images for the whole area when zoomed far away, prepare still the combined shapefile with covered by -images removed.  Then make two scale dependent layers under the same group.  Read 2 m data when scale is between 200k and 100k and switch then to the combined layer.
	 
	There may be better ways, this is the best I have invented this far for our own use.
	 
	Take care of creating enough many low-resolution layers with gdaladdo for far away zooms.  Even then if Mapserver must open more than 20-40 image files for filling the screen I would consider making a separate overview image by mosaicing the whole area together with something like 20-50-100 meter pixel size and using that image in the same image group.
	 
	-Jukka Rahkonen-


________________________________

		Lähettäjä: mapserver-users-bounces at lists.osgeo.org [mailto:mapserver-users-bounces at lists.osgeo.org] Puolesta Liang Jian
		Lähetetty: 16. marraskuuta 2009 13:02
		Vastaanottaja: mapserver-users at lists.osgeo.org
		Aihe: [mapserver-users] performance problem with ortho images
		
		
		
		Hi all,
		 
		I have a problem with mapserver handling ortho images. The situation is we have 2m data cover the whole country and 20-50cm data cover most cities. We setup mapfiles as 2m data shows between scale 200k to 500 and 20-50cm data show between scale 100k to 500. Problem is when requirst data in the city ranges, mapserver first read data from 20-50cm dataset and then read from 2m dataset which is unnecessary as the 20-50cm data is enough and the best solution is when there is no 20-50cm data, 2m data should be used instead. 
		 
		Do you have any idea how to avoid the unnecessary reading? 
		 
		Any suggestions will be welcome!
		 
		Thanks
		 
		/Liang

________________________________

		Confidentiality Notice: This e-mail transmission may contain confidential or legally privileged information that is intended only
		for the individual or entity named in the e-mail address. If you are not the intended recipient, you are hereby notified that any
		disclosure, copying, distribution, or reliance upon the contents of this e-mail is strictly prohibited and may be unlawful. If you have received this e-mail in error, please notify the sender immediately by return e-mail and delete all copies of this message.
		


________________________________

	Confidentiality Notice: This e-mail transmission may contain confidential or legally privileged information that is intended only
	for the individual or entity named in the e-mail address. If you are not the intended recipient, you are hereby notified that any
	disclosure, copying, distribution, or reliance upon the contents of this e-mail is strictly prohibited and may be unlawful. If you have received this e-mail in error, please notify the sender immediately by return e-mail and delete all copies of this message.
	

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/mapserver-users/attachments/20091116/fa957179/attachment.html