[4eyes] Fwd: ISPRS benchmark on urban object detection and 3D building reconstruction
Steffen Gauglitz
sgauglitz at cs.ucsb.edu
Mon Jan 14 16:02:41 PST 2013
This benchmark and dataset may of interest to ppl working on 3D
reconstruction etc ...
Cheers,
Steffen
-------- Original Message --------
Subject: ISPRS benchmark on urban object detection and 3D building
reconstruction
Date: Mon, 14 Jan 2013 23:59:02 +0000
From: Bob Fisher <rbf at inf.ed.ac.uk>
To: sgauglitz at cs.ucsb.edu
Dear colleagues,
[Please contact Markus Gerke <m.gerke at utwente.nl> about this information]
This e-mail is intended to give you an update on new developments in the
context of the ISPRS
benchmark on urban object detection and 3D building reconstruction,
being conducted by the
working group III/4 "3D Scene Analysis" of the ISPRS
(http://www.commission3.isprs.org/wg4/
[www.commission3.isprs.org]).
1) On the benchmark
The automated extraction of urban objects from data acquired by airborne
sensors has been an
important topic of research in photogrammetry for at least two decades.
One observation that
can be made is that most of the papers dealing with urban object
extraction focus on a single
object class, e.g. buildings, roads or, less frequently, trees.
Typically, results are
published for a few test sites available to the authors.
The aim of this benchmark test is to provide state-of-the-art data sets
which can be used by
interested researchers in order to test own methods and algorithms on
urban object
classification and building reconstruction.
First results are available online and were also published at the ISPRS
congress in Melbourne,
last August; see the web site the benchmark for details and for a data
request form:
http://www.itc.nl/ISPRS_WGIII4/tests_datasets.html [www.itc.nl],
You are strongly encouraged to submit results of your latest methods.
All submitted results
will be continuously evaluated.
2) New data:
For Vaihingen, new input data are available, namely a digital surface
model (DSM) from image
matching and a true orthophoto mosaic generated on the basis of this
DSM. These data should
make it easier for participants to generate results for urban object
detection. We also
encourage participants who have already submitted results based on the
DSM from ALS to apply
their methods also to the DSM from image matching, in order to be able
to assess the benefits
of using ALS data compared to a purely image-based work-flow.
For Toronto, the orientation data of the images (which contained an
error) were updated.
Furthermore, we have made available more images.
3) Relaxed conditions of use:
For Vaihingen, the restriction that the data may only be used in the
context of the ISPRS
benchmark has been dropped. Thus, the data may be used for any research
purposes, and results
achieved on the data may be published in any scientific journal or at
any scientific
conference.
For Toronto, the NDA is no longer required. That is, the data may be
used for research in the
context of the ISPRS benchmark.
4) Workshop at the CRV conference
We are also very pleased to announce that the ISPRS WGIII/4 committee
will organize a joint
workshop of Virtual City Modelling (VCM) 2013 in conjunction with the
Tenth Conference on
Computer and Robotic Vision (CRV) 2013 on the 28th - 31st of May 2013 at
the University of
Regina, Regina, Canada. The main purpose of organizing this workshop is
to gather the latest
multidisciplinary research works in virtual city modelling from all of
the fields in
photogrammetry, remote sensing, computer vision, machine learning, and
computer graphics. It
will be a great opportunity to share the research of the participants in
the ISPRS benchmark
test project. We strongly encourage all participants to attend this
workshop, as it will offer
a chance to exchange ideas, identify current and future research
problems, and engage in deep
discussion with researchers in diverse fields. More detailed information
on this workshop can
be found on the following website: http://www.geoict.yorku.ca/vcm2013/
[www.geoict.yorku.ca]
5) Theme Issue of the ISPRS Journal of Photogrammetry and Remote Sensing
There will be a theme issue of the ISPRS Journal of Photogrammetry and
Remote Sensing on "Urban
Object Detection and 3D Building Reconstruction"
(see http://www.itc.nl/ISPRS_WGIII4/docs/cfp_urban_object_extraction.pdf
[www.itc.nl]). We
encourage all test participants to submit papers to this theme issue;
deadline is the 31st of
May, 2013.
Best regards,
Franz Rottensteiner, Gunho Sohn, Markus Gerke, Jan D. Wegner
--
http://www.itc.nl/about_itc/whoiswho/viewone.asp?Markus%20Gerke [www.itc.nl]
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
More information about the Ilab-users
mailing list