[GRASS-dev] Introduction and GSoC

[Eric Momsen]

I am reading more about image classification (from
http://grass.osgeo.org/wiki/GRASS_SoC_Ideas ):

4. Implement image segmentation algorithms and tools
5. Implement region-based classification
6. Implement hierarchical classification tools (e.g. being able to
create a large class "forest", with subclasses of different types of
forests)

I see Hamish is interested in mentoring the parallelization portion of
that list.  Are these other ideas orphans, or is someone available
that could discuss the background and needs of the community around
these ideas (and/or mentor...)  Thanks!

(And uDig conversation continued below...)

On Fri, Mar 23, 2012 at 5:47 PM, Wolf Bergenheim
<wolf+grass at bergenheim.net> wrote:
> On Fri, Mar 23, 2012 at 22:04, Eric Momsen <eric.momsen at gmail.com> wrote:
>>
>> On Fri, Mar 23, 2012 at 2:41 PM, Wolf Bergenheim
>> <wolf+grass at bergenheim.net> wrote:
>> > Hi Eric!
>> >
>> > It might help us in trying to classify you if you could tell a bit about
>> > your programming experience. What languages are you comfortable with? I
>> > see
>>
>> So far I have been using mostly java.  I wrote a small implementation
>> for a pattern mining algorithm in R, and will be doing a little prolog
>> this semester.  Years ago I helped set up a data transfer/validation
>> system ( Oracle database <-> visual basic <-> website), but those
>> skills are pretty rusty now.  I certainly expect to learn more
>> languages.
>>
>
> Hmm If you are most strong in Java I'd recommend looking at uDig
> (http://udig.refractions.net/confluence/display/HACK/Summer+of+Code), since
> GRASS is C and Python. It might be a better match.
>

I'm happy to learn Python, it seems the software goal is more
important then the language that will be used.

My impression is that the goal is to provide an easy way for the user
to *manually* explore data from any available source.  The
visualization and intuitive interface look great.  But am I correct
that for doing statistics and number crunching, I should be focusing
on GRASS and R?

>>
>> > that you are interested in data mining, which is unfortunately something
>> > I
>> > know very little about... Did you see any ideas which interested you?
>> > You
>>
>> The first project for animating raster/(and possibly vector) time
>> series data was interesting, as well as a number of the imagery
>> topics.
>>
>
> Cool, You should probably think about more than one topic and maybe prepare
> several applications.
>
>>
>> > can of course also come up with your own project. Is there for instance
>> > some
>> > data mining module missing from GRASS? That could be interesting,
>> > especially
>> > if you prepare a good solid proposal for it.
>>
>> I was actually hoping to get farther along with my own research to be
>> able to (selfishly) propose something directly helpful to my work.  (I
>> do understand the goal is the code, but if I could meet both goals...)
>>  So far I seem to be stuck on the pre-processing side of things.  I
>> will touch base with my adviser again next week to see if we can come
>> up with something interesting to propose.  If we do come up with
>> something, should I sketch the idea here first?  Is someone available
>> to mentor?
>>
>
> There might be someone willing to mentor. That remains to be seen. Would
> your adviser be interested in helping to co-mentor you, along with a GRASS
> mentor? It would definitely be a good idea to discuss your ideas here first,
> it will improve the chances of getting selected.
>
>  --Wolf
>