[gdal-dev] [RFC] [GDAL] Idea for GSoC, 2014

Fri Mar 14 06:28:36 PDT 2014

Hello everyone

Continuing the previous discussion, I would like to propose something and
the community's suggestions are welcomed/needed. I can understand that this
thread is a little old, so let me remind you that its regarding the
automatic geo-referencer idea. The idea is also proposed on the GDAL ideas
page (http://trac.osgeo.org/gdal/wiki/SummerOfCode).

Based on the previous discussions, what came out was that we can improve
the current implementation of SIMPLE SURF in GDAL which was developed as a
part of 2012 GSOC GDAL Correlator project, to support *large data* and *multi
spectral imagery*. And then apply this *modified* algorithm for the
geo-reference purposes. Now I have been in touch with Chaitanya, who is
willing to mentor this project, and there are some things on which we would
like to know community's suggestions/response.

There are basically two things that can be done regarding this project:

1. As mentioned above, we can modify the SIMPLE SURF algorithm and make it
much better for the geo-reference purposes. Already, a lot had been
discussed on this and we have a fairly good idea about what is to be done.

2. One more thing that can be done is that we can implement BRISK
algorithm[1] instead of SURF along with the FLANN matcher for this purpose.
What advantages this thing offers is that it is fairly fast and gives
comparable outputs along with that it works well with fairly large data
sets. So we do not need to segment the imagery as we would have done in the
case of SURF. Also added to this, this algorithm also has no patent issues.
We had a lot of problem regarding patent issues in SIFT/SURF and we
discussed them at length on the mailing list as well.

One thing that I fell can be done is that  two proposal can be written, one
for each and then community can decide accordingly which one is more
useful. Or we can decide it here itself..?

Kindly provide your valuable comments and suggestion..

With Regards,

Kshitij Kansal

Lab For Spatial Informatics,

IIIT Hyderabad

1. http://www.robots.ox.ac.uk/~vgg/rg/papers/brisk.pdf

On Thu, Feb 13, 2014 at 2:23 PM, Kshitij Kansal <kansal.k at gmail.com> wrote:

> Hello Everyone
>
> I have updated the Summer Of Code Ideas Page for GDAL. I have Introduced
> the above idea in that page also. Mr. Chaitanya Kumar is willing to mentor
> the project.
>
> Please look into this.
>
> Suggestions and Comments are welcomed.
>
> Regards,
>
> Kshitij Kansal
>
> Lab For Spatial Informatics,
>
> IIIT Hyderabad
>
>
>
> On Sat, Feb 8, 2014 at 2:07 PM, Dmitriy Baryshnikov <bishop.dev at gmail.com>wrote:
>
>>  Hi,
>>
>> Even note that you cannot get the code from this projects and merge it to
>> GDAL, but you  free to develop you own implementation of this algorithms as
>> path of GDAL.
>> This is a same situation with correlator. Andrew cannot get code from
>> OpenCV and GRASS as incompatible licenses, so the own (simple)
>> implementation was provided.
>> One can change this implementation to be more advanced: support large
>> datasets, more bands, etc. Everybody  welcome to do it.
>>
>> Best regards,
>>     Dmitry
>>
>> 08.02.2014 2:15, Kshitij Kansal пишет:
>>
>> Sir
>>
>>  I am providing the links for both the algorithms that I talked. I am
>> new to this licence issues thing so I would be highly grateful if you can
>> clarify doubts and if can proceed on working for this idea.
>>
>>  http://www.cs.ubc.ca/~lowe/keypoints/ (SIFT)
>>  http://www.ipol.im/pub/art/2011/my-asift/ (ASIFT) [here source code is
>> provided. It contains license ]
>>  http://www.cmap.polytechnique.fr/~yu/research/ASIFT/demo.html (Demo
>> Page for ASIFT)
>>
>>  The SIFT algorithm in itself is patented. They only provide binaries. I
>> am not sure If we can use those binaries or not.
>> As far as ASIFT is concerned, i think its implementation is only
>> patented.
>>
>>  Kindly clarify this.
>>
>>  Thanking You
>> With Regards,
>>
>>  Kshitij Kansal
>>
>> Lab For Spatial Informatics,
>>
>> IIIT Hyderabad
>>
>>
>>
>> On Sat, Feb 8, 2014 at 2:23 AM, Even Rouault <
>> even.rouault at mines-paris.org> wrote:
>>
>>> Selon Kshitij Kansal <kansal.k at gmail.com>:
>>>
>>> Kshitij,
>>>
>>> I'm surprised that you mention licenses
>>> for algorithms. What are your sources for that?
>>> Only *implementations* can be licensed
>>> not algorithms themselves. So if you
>>> develop your own implementation you are free
>>> to select the license you wish. There
>>> might be issues linked to potential pattents
>>> in some countries but that is acceptable in my opinion
>>> as software parents are not valid in all contries
>>> so mentionning potential issues with them
>>> in the documentation is sufficient .
>>>
>>> > Hello
>>> >
>>> > Based on all suggestion
>>>  ins and comments above, I have come up with one thing.
>>> >
>>> > The most important thing that could have been a road block for this
>>> project
>>> > was licence issues which was pointed by Even Rouault.
>>> >
>>> > I am interested in implementing the SIFT or ASIFT algorithms for
>>> automatic
>>> > geo-referencing which I proposed in above mails. I am interested in
>>> making
>>> > a separate tool which will be completely based on the GDAL and
>>> maintained
>>> > by GDAL community. I looked into the GDAL-correlator idea which was
>>> also
>>>  > implemented as a part of GSoC. That idea uses only *SIMPLE SURF
>>> *algorithm
>>> > for control point detection and is only *limited to small size three
>>> band
>>> > imagery(RGB)*.
>>> >
>>> > I am interested in developing something that is fully automated and
>>> works
>>> > on large multi band imagery.But the above two algorithms that I am
>>> planning
>>> > to use come up with the licence that does not allows the free use for
>>>  > commercial purposes. The free use of above algorithm are restricted
>>> to *"non
>>> > profit research and non profit educational purposes"*(in case of
>>> > ASIFT) and *"research
>>>  > purposes only" *(in case of SIFT).
>>>  >
>>> > This means that the final product as a result of this product will
>>> also be
>>> > bounded by the above licences. It can not go into GDAL's main
>>> distribution
>>> > but can be used as a separate utility of GDAL for non-commercial uses
>>> only.
>>> > Basically the commercial use of the tool won't be allowed without
>>> approval
>>> > from the concerned people.
>>> >
>>> > Your views and suggestions are highly appreciated.
>>> >
>>> > Regards,
>>> >
>>> > Kshitij Kansal
>>> >
>>> > Lab For Spatial Informatics,
>>> >
>>> > IIIT Hyderabad
>>> >
>>> >
>>> >
>>> > On Thu, Jan 30, 2014 at 12:25 PM, Kshitij Kansal <kansal.k at gmail.com>
>>> wrote:
>>> >
>>> > > Hello
>>> > >
>>> > > @Jukka Rahkonen: The OSSIM project (the link you provided) is more of
>>> > > image orthorectification. Although
>>> > > they are doing image co-registration but its as one of the steps of
>>> > > orthorectification(I could only understand this from the manual).
>>> > > Also I am not sure of the techniques they are using for this
>>> purposes.
>>> > > (Its written corner point detection but How?)
>>> > > I am new to OSSIM, so not aware of this thing's functionality and
>>> > > accuracy. I am in the process of estimating it.
>>> > >
>>> > > Thank you for pointing this out.
>>> > >
>>> > > Coming to the Frank Warmerdam's blog (http://fwarmerdam.blogspot.fi/
>>> ).
>>> > > Here the author is talking about writing a code for converting the
>>> GCP's
>>> > > into RPC's. But the next question comes is where is he getting GCP's
>>> from.
>>> > > Now this is the part where I aim to work on and improve.
>>> > >
>>> > > Given two images, if we can automatically extract GCP's from them
>>> then we
>>> > > are in a way speeding up and automating this whole process. Once we
>>> have
>>> > > GCP, there are a lot of things that we can do from them like
>>> > > geo-referencing(which I plan on doing), image stitching etc.
>>> > >
>>>  > > *Also you talked about a "more light weighted system". Can you
>>> please
>>> > > elaborate on this?* I could not understand what you actually meant
>>> from
>>> > > that.
>>> > >
>>> > > If we look into the last year idea's page (
>>> > > http://trac.osgeo.org/gdal/wiki/SummerOfCode), there is one idea
>>> that is
>>> > > proposed regarding the "Raster/Vector geo-referencer on the Web". I
>>> want to
>>> > > modify this idea a little bit and want to automate the whole
>>> process. As
>>> > > far as the automation is concerned, we can develop GDAL functions for
>>> > > that(which will be completely re-usable) and then developing a Web
>>> based
>>> > > geo-referencer would only require calling those functions.
>>> > >
>>>  > > If we do this thing,* two objectives will be fulfilled*.* First*,
>>> GDAL
>>> > > will get a web based geo-referencer. *Secondly*, communities that use
>>>  > > GDAL and have not developed these kind of geo-referencer(Like
>>> QGIS) can
>>> > > directly use the functions developed and then build on that.
>>> > >
>>> > > Suggestions and comments are welcomed.
>>> > >
>>> > > Thank you for your propositions.
>>> > >
>>> > > Regards,
>>> > >
>>> > >  Kshitij Kansal
>>> > >
>>> > > Lab For Spatial Informatics,
>>> > >
>>> > > IIIT Hyderabad
>>> > >
>>> > >
>>> > >
>>> > > On Wed, Jan 29, 2014 at 1:27 PM, Jukka Rahkonen <
>>> jukka.rahkonen at mmmtike.fi
>>> > > > wrote:
>>> > >
>>> > >> Kshitij Kansal <kansal.k <at> gmail.com> writes:
>>> > >>
>>> > >> >
>>> > >> > Even,
>>> > >> > Thank you for pointing out this issue. I will keep this in mind.
>>> > >> >
>>> > >> > I will look into that last year project and try to understand the
>>> > >> implementation.
>>> > >> >
>>> > >>
>>> > >> > More suggestions and comments are always welcomed.
>>> > >>
>>> > >> Hi,
>>> > >>
>>> > >> Another OSGeo project OSSIM has implemented an automatic image to
>>> image
>>> > >> rectification. Isn't that a bit alike your plan?
>>> > >>
>>> http://download.osgeo.org/ossim/docs/OSSIM_Coregistration_UserManual.pdf
>>> > >>
>>> > >> Frank Warmerdam seems to entertain himself with something related
>>> > >> http://fwarmerdam.blogspot.fi/
>>> > >>
>>> > >> This article from 2013 may also give some inspiration
>>> > >>
>>> > >>
>>> >
>>>
>>> http://www.academia.edu/4388853/OrientAL_-_Automatic_geo-referencing_and_ortho-rectification_of_archaeological_aerial_photographs
>>> > >>
>>> > >> All these three aim to very high quality and I am sure that there
>>> is also
>>> > >> need for a more light-weight system.
>>> > >>
>>> > >> -Jukka Rahkonen-
>>> > >>
>>> > >> _______________________________________________
>>> > >> gdal-dev mailing list
>>> > >> gdal-dev at lists.osgeo.org
>>> > >> http://lists.osgeo.org/mailman/listinfo/gdal-dev
>>> > >>
>>> > >
>>> > >
>>> >
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>> _______________________________________________
>> gdal-dev mailing listgdal-dev at lists.osgeo.orghttp://lists.osgeo.org/mailman/listinfo/gdal-dev
>>
>>
>>
>> _______________________________________________
>> gdal-dev mailing list
>> gdal-dev at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/gdal-dev
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20140314/184128fb/attachment-0001.html>