global mapping issues with orthographic projection

Frank Warmerdam fwarmerdam at GMAIL.COM
Sat Apr 23 10:54:17 EDT 2005

On 4/22/05, Ken Tanaka <ken.tanaka at> wrote:
> At the moment we are having some troubles with a "starburst" of stray
> feature lines that can be seen in this URL:


I don't *think* you are running into any of the traditional MMS world mapping
issues yet. 

The "starburst" effect is quite odd.  It looks like some vertices are "at" 0,0
for some reason.  I would suggest:

 o trying using an OGR connection type in you mapfile, to see if there is some
   quirk reading your shapefiles with the builtin mapserver code. 

 o If that produces the same results, inspect your dataset with ogrinfo and 
   see if there is anything odd about the vertices.  Ideally reviewing a 
   polygon you know is causing problems like Maine. 

Other points: 
 o As Sean notes, it is hard to believe your data is really in an orthographic
    projection.  If your shapefile is really in geographic latlong
then you should
    declare it appropriately in your mapfile (+proj=latlong +datum=NAD83). 

 o You declare your state and country layer projections without any lat_0, 
    or lon_0 values.  This is dangerous because internal defaults will be
    used and these can be hard to predict.  I imagine MapServer is trying
    to reproject the vectors between two ortho projections but I doubt that
    is what you really want to happen. 
> Our current approach is to generate the lines with MMS, and then use a
> separate step to merge the data (satellite image) with the outlines image,
> and produce the composite image seen in the above URL.

Is there a reason you don't have MapServer taking care of the image data
as well? 

Best regards,
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at
light and sound - activate the windows |
and watch the world go round - Rush    | Geospatial Programmer for Rent

More information about the mapserver-users mailing list