| Re: Gridding Satellite Data [message #77330] |
Tue, 23 August 2011 11:01  |
R.G.Stockwell
Messages: 163
Registered: October 2004
|
Senior Member |
|
|
>>
>>
>> "David Fanning" wrote in message
>> news:MPG.28bd5db2d56df65c989961@news.giganews.com...
>>
>> Folks,
>>
>> The typical sequence of events when you are contemplating
>> asking a question that requires remotely sensed data, say
>> from a satellite, is to start with some region you are
>> interested in. At the National Snow and Ice Data Center,
>> for example, this region is described as a rectangular
>> area in an XY or Cartesian coordinate system, with a
>> particular spatial resolution, say 25km. AT NSIDC, this
>> area is described by a "gpd" (grid parameter definition)
>> file. A map projection, of course, is a required element
>> in setting up the Cartesian coordinate system.
>>
>> The next step is to gather satellite data for the various
>> instruments (e.g., MODIS, MISR, etc) that happen to fly
>> over this area, and to grid the data into this rectangular
>> study area. At NSIDC, for example, a C program named mapx is
>> used to do this gridding. Mapx allows you to do the gridding
>> in several different ways, with different types of pixel
>> weighting, etc.
>>
>> This part of the process I have never been able to do
>> successfully in IDL.
>>
>> So, here is my question. What sort of software do YOU
>> use to perform this operation of taking available
>> satellite data and preparing it for a remote sensing
>> study?
>>
>> I am looking for what I hope will be a machine-independent
>> solution to this problem. In other words, I am hoping
>> for a solution I can run on a Windows machine. :-)
>>
>> Cheers,
>>
>> David
>>
>>
>> --
>> David Fanning, Ph.D.
>> Fanning Software Consulting, Inc.
>> Coyote's Guide to IDL Programming: http://www.idlcoyote.com/
>> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
>
>
> "David Fanning" wrote in message
> news:MPG.28bd5db2d56df65c989961@news.giganews.com...
>
> Folks,
>
> The typical sequence of events when you are contemplating
> asking a question that requires remotely sensed data, say
> from a satellite, is to start with some region you are
> interested in. At the National Snow and Ice Data Center,
> for example, this region is described as a rectangular
> area in an XY or Cartesian coordinate system, with a
> particular spatial resolution, say 25km. AT NSIDC, this
> area is described by a "gpd" (grid parameter definition)
> file. A map projection, of course, is a required element
> in setting up the Cartesian coordinate system.
>
> The next step is to gather satellite data for the various
> instruments (e.g., MODIS, MISR, etc) that happen to fly
> over this area, and to grid the data into this rectangular
> study area. At NSIDC, for example, a C program named mapx is
> used to do this gridding. Mapx allows you to do the gridding
> in several different ways, with different types of pixel
> weighting, etc.
>
> This part of the process I have never been able to do
> successfully in IDL.
>
> So, here is my question. What sort of software do YOU
> use to perform this operation of taking available
> satellite data and preparing it for a remote sensing
> study?
>
> I am looking for what I hope will be a machine-independent
> solution to this problem. In other words, I am hoping
> for a solution I can run on a Windows machine. :-)
>
> Cheers,
>
> David
>
>
> --
> David Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote's Guide to IDL Programming: http://www.idlcoyote.com/
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
The implications of the sampling, and what the desired analysis is, have a
great influence on how to go about the next step. Simple gridding is often
not a sufficient approach, due to the aliasing (etc) of the data if one does
not have a geostationary satellite (and many many measurements are from
lower earth orbit in precessing polar orbits).
So, I would give 2 answers.
1) I have a LOESS smoothing/interpolation code that has been used in the
past for satellite data. It does local 3D least squares fitting to create a
uniformly sampled dataset of measurements (lon,lat,time). Adding height
would not be a problem. Depending on the goals of the analysis, it might be
useful.
2) hire a post-doc and perhaps look into employing the Salby Method to the
data (if appropriate).
Bonus answer 3) it depends. On the measurement, the orbit, the sampling
characteristics, the goals of the analysis, and therefore a general solution
might not be possible.
cheers,
bob
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
