Hi all, we have a program we announced a year or so ago called "starspan"
(starspan.casil.ucdavis.edu) -- the purpose of this program is to query
raster data based on vector input(s) and produce a number of outputs
(including a .csv file with the vector database + raster data fused
together). This algorithm is FAR faster than equivalent algorithms (e.g.
ESRI's zonal stats) and was designed specifically with large datasets (and
multiple raster and vector datasets) in mind and for import into stats
packages.
We're hoping to get some feedback or even an offer of help to roll this into
a future grass release -- if anyone is interested please contact me (take a
look at the website so you can see what we are trying to accomplish).
Ideally, we'd like someone familiar with grass's image analysis algorithms
(e.g. classification), since one of the major applications of this program
is to speed up the process of initial classification -- the idea is you
extract the spectral data associated with "field" data (known classes,
perhaps collected with a GPS), perform the classifications/accuracy
assessments only on this subset (which will be far smaller than the entire
image), and once the correct classification is decided on, only then is it
applied to the entire image(s). We did this with ENVI quite a bit, and it
REALLY sped up how we analyzed data.
Much of the algorithm is finished, we mostly need help getting the program
grass-ready, and cleaning up some of the interface.
Please let me know! I use AIM and MSN (accounts below) if you would prefer
instant messaging, otherwise please email me!
--j
--
Jonathan A. Greenberg, PhD
NRC Research Associate
NASA Ames Research Center
MS 242-4
Moffett Field, CA 94035-1000
Office: 650-604-5896
Cell: 415-794-5043
AIM: jgrn3007
MSN: jgrn3007@hotmail.com
On Wednesday 15 February 2006 04:11 pm, Jonathan Greenberg wrote:
Hi all, we have a program we announced a year or so ago called "starspan"
(starspan.casil.ucdavis.edu) -- the purpose of this program is to query
raster data based on vector input(s) and produce a number of outputs
(including a .csv file with the vector database + raster data fused
together). This algorithm is FAR faster than equivalent algorithms (e.g.
ESRI's zonal stats) and was designed specifically with large datasets (and
multiple raster and vector datasets) in mind and for import into stats
packages.
We're hoping to get some feedback or even an offer of help to roll this
into a future grass release -- if anyone is interested please contact me
(take a look at the website so you can see what we are trying to
accomplish). Ideally, we'd like someone familiar with grass's image
analysis algorithms (e.g. classification), since one of the major
applications of this program is to speed up the process of initial
classification -- the idea is you extract the spectral data associated with
"field" data (known classes, perhaps collected with a GPS), perform the
classifications/accuracy assessments only on this subset (which will be far
smaller than the entire image), and once the correct classification is
decided on, only then is it applied to the entire image(s). We did this
with ENVI quite a bit, and it REALLY sped up how we analyzed data.
Much of the algorithm is finished, we mostly need help getting the program
grass-ready, and cleaning up some of the interface.
Please let me know! I use AIM and MSN (accounts below) if you would prefer
instant messaging, otherwise please email me!
--j
--
Jonathan A. Greenberg, PhD
NRC Research Associate
NASA Ames Research Center
MS 242-4
Moffett Field, CA 94035-1000
Office: 650-604-5896
Cell: 415-794-5043
AIM: jgrn3007
MSN: jgrn3007@hotmail.com
Jonathan and other GRASS users,
I can certainly vouch for STARSPAN's speed and flexibility. I have been using
it (with some great help from the author and current maintainer) in my
research for about a month now. I am glad to hear that Jonathan is interested
in making a tighter coupling between this tool and GRASS, something I have
often thought about doing myself (yeah right!). This message is very timely,
as I have recently gone through a review of the various ways in GRASS to
accomplish this same task. This tool would be a great enhancement to the
stats tools currently in the GRASS toolset.
I would be willing to spend some time working on this, as it is directly
realted to my research- however oversight by a more capable programmer would
be a must.
Cheers,
--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341