[Geoserver-users] Sinusoidal Projection

GeoServer GeoServer User
1

Hi,
Sorry to be bomdarding the list with questions.

I have a series of coverages with the following projection (WKT):

PROJCS["User_Defined_Sinusoidal", GEOGCS["GCS_User_Defined",
DATUM["User_Defined", SPHEROID["User_Defi
ned_Spheroid",6371007.181,0.0]], PRIMEM["Greenwich",0.0],
UNIT["Degree",0.0174532925199433]], PROJECTI
ON["Sinusoidal"], PARAMETER["False_Easting",0.0],
PARAMETER["False_Northing",0.0], PARAMETER["Central_
Meridian",0.0], UNIT["Meter",1.0]]

I added this projection to data_dir/user_projections/epsg.properties
file.

After restarting the server I went to list the available projections and
my projection does not appear. However, the server spits out the
following error message:

63880 [btpool0-6] INFO /geoserver - Initializing Spring FrameworkServlet
'dispatcher'
25 Jun 11:20:06 WARN [action.SrsHelpAction] - Issues converting
EPSG:100005. Error in "PROJECTION": No
transform for classification "Sinusoidal". Stack trace included at FINE
logging level
25 Jun 11:20:10 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:10 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:14 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:14 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.

It is odd that GeoServer can't handle this projection. I can reproject
the datasets without much trouble using GDAL.

The data we wish to provide is for scientific purposes so we don't want
to reproject it and risk messing up the data values.

Thank you for any information.
Regards,
Craig Dillabaugh

2

Dillabough, Craig ha scritto:

Hi,
Sorry to be bomdarding the list with questions.

I have a series of coverages with the following projection (WKT):

PROJCS["User_Defined_Sinusoidal", GEOGCS["GCS_User_Defined",
DATUM["User_Defined", SPHEROID["User_Defi
ned_Spheroid",6371007.181,0.0]], PRIMEM["Greenwich",0.0],
UNIT["Degree",0.0174532925199433]], PROJECTI
ON["Sinusoidal"], PARAMETER["False_Easting",0.0],
PARAMETER["False_Northing",0.0], PARAMETER["Central_
Meridian",0.0], UNIT["Meter",1.0]]

I added this projection to data_dir/user_projections/epsg.properties
file.

After restarting the server I went to list the available projections and
my projection does not appear. However, the server spits out the
following error message:

63880 [btpool0-6] INFO /geoserver - Initializing Spring FrameworkServlet
'dispatcher'
25 Jun 11:20:06 WARN [action.SrsHelpAction] - Issues converting
EPSG:100005. Error in "PROJECTION": No
transform for classification "Sinusoidal". Stack trace included at FINE
logging level
25 Jun 11:20:10 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:10 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:14 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:14 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.

It is odd that GeoServer can't handle this projection. I can reproject
the datasets without much trouble using GDAL.

Why odd? As far as I can see none of the official EPSG codes uses it.
It can be implemented thought, if you have development abilities I can
point you to other projections as examples or someone can implement it
for you as a consulting job, should not be that difficult, the math is
simple:
http://en.wikipedia.org/wiki/Sinusoidal_projection

Looking at spatial-reference.org it would seem the projection is used
mostly for mapping other planets imagery (with a notable exception with
an ESRI only african CRS and a earth wise one, none of them being official EPSG codes thought)
http://www.spatialreference.org/ref/search=sinusoidal

None of the SRS listed by postgis uses sinusoidal either (thought I
know GDAL is capable of handling it).

Cheers
Andrea

3

Hi Andrea,
Thanks again.

Sinusoidal is the projection the USGS provides Modis data in. It is
rather bizarre and ugly looking (for Canada at least), but we want to
fiddle with the data as little as possible.

As you mentioned, that math does look pretty easy. If you could please
point me to some examples of how reprojections are done I will see if it
is something I can handle.

Regards,
Craig

-----Original Message-----
From: Andrea Aime [mailto:aaime@anonymised.com]
Sent: Friday, June 26, 2009 9:04 AM
To: Dillabough, Craig
Cc: geoserver-users@lists.sourceforge.net
Subject: Re: [Geoserver-users] Sinusoidal Projection

Dillabough, Craig ha scritto:

Hi,
Sorry to be bomdarding the list with questions.

I have a series of coverages with the following projection (WKT):

PROJCS["User_Defined_Sinusoidal", GEOGCS["GCS_User_Defined",
DATUM["User_Defined", SPHEROID["User_Defi
ned_Spheroid",6371007.181,0.0]], PRIMEM["Greenwich",0.0],
UNIT["Degree",0.0174532925199433]], PROJECTI
ON["Sinusoidal"], PARAMETER["False_Easting",0.0],
PARAMETER["False_Northing",0.0], PARAMETER["Central_
Meridian",0.0], UNIT["Meter",1.0]]

I added this projection to data_dir/user_projections/epsg.properties
file.

After restarting the server I went to list the available projections

and

my projection does not appear. However, the server spits out the
following error message:

63880 [btpool0-6] INFO /geoserver - Initializing Spring

FrameworkServlet

'dispatcher'
25 Jun 11:20:06 WARN [action.SrsHelpAction] - Issues converting
EPSG:100005. Error in "PROJECTION": No
transform for classification "Sinusoidal". Stack trace included at

FINE

logging level
25 Jun 11:20:10 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:10 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:14 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.
25 Jun 11:20:14 WARN [referencing.factory] - Ambiguity between inverse
flattening and semi minor axis
length. Using inverse flattening.

It is odd that GeoServer can't handle this projection. I can

reproject

the datasets without much trouble using GDAL.

Why odd? As far as I can see none of the official EPSG codes uses it.
It can be implemented thought, if you have development abilities I can
point you to other projections as examples or someone can implement it
for you as a consulting job, should not be that difficult, the math is
simple:
http://en.wikipedia.org/wiki/Sinusoidal_projection

Looking at spatial-reference.org it would seem the projection is used
mostly for mapping other planets imagery (with a notable exception with
an ESRI only african CRS and a earth wise one, none of them being
official EPSG codes thought)
http://www.spatialreference.org/ref/search=sinusoidal

None of the SRS listed by postgis uses sinusoidal either (thought I
know GDAL is capable of handling it).

Cheers
Andrea

4

Dillabough, Craig ha scritto:

Hi Andrea,
Thanks again.

Sinusoidal is the projection the USGS provides Modis data in. It is
rather bizarre and ugly looking (for Canada at least), but we want to
fiddle with the data as little as possible.

As you mentioned, that math does look pretty easy. If you could please
point me to some examples of how reprojections are done I will see if it
is something I can handle.

There is this tutorial that you can follow:
http://docs.codehaus.org/display/GEOTDOC/How+to+add+new+projections

It may be a little outdated but should give you a good starting point.
Besides that download the GeoTools sources and look into the
modules/library/referencing module where all of the other projections
are implemented.

Cheers
Andrea