- Archival media readability and integrity: How long can current media
be expected to survive, and will it be readable after two or three
generations of new media? How can we insure that critical data survive
this process of technological evolution with integrity?
- Archival site stability: How can we insure that archival sites (especially
those referenced in the archival literature) remain accessible for
reasonable periods of time? How long a time period should be considered
minimal for public access?
- Archival site size vs access bandwidth trends: Do the Moore's law
exponents for storage media and internet bandwidth suggest a trend
that capacity will eventually exceed ability to access data in reasonable
time? If so, should compression, filtering, and estimation/detection
processing algorithms along with data originator, data format, calibration
issues, etc. which help researchers to experiment with raw data be
located at the archival site?
- Data visualization: How can we provide on-line visualization tools
that can assist user in identifying meaningful data subsets within
large sets?
- A visual archive of satellite data. We need to compile thumbnail
of thousands of images, sorted out according to a selected criterion,
and provide them along with a browser on CDs to users. The browser
that was displayed in IGARSS'02 (in the exhibition area) will serve
this purpose. An important theme for this research priority is the
arrangement of archive data in a user-selected sequence (or hierarchy)
in order to facilitate selecting coincident and co-located data. This
is particularly important as work on data fusion is growing rapidly
and the synergistic use of multi-platform data is so much encouraged.
This proposal is for an off-line browsing tool. As for an on-line
tool, it is certainly even more important and needed. It can be identified
as a priority but any reasonable solution to bring on-line data from
different sources into one system will be technically challenging
and expensive.
- As we intend to provide data and information to the application
communities, it might be useful to identify products of most socio/economic
values along with the most recent and robust algorithms to generate
those products. The list should be updated regularly as sensors and
applications continue to evolve. We want, for example to make the
user community aware that DEM can be produced accurately from SAR
interferometry, or sea ice surface temperature has become available
from IR channels, or a new vegetation index has been produced from
MODIS data and so on. This research priority may come under item "distributed
resources" in Roger's paper
(IGARSS02).
- Data compression. Recently, there has been much work on this topic.
Which kind of compression do we need for DAD purposes? This point
is also related to the network infrastructure (and bandwidth) through
which users are going to access the archives. Can we use lossless
compression? Do we need near-lossless or even lossy compression, and
in the affirmative, what are the quality issues related to data alteration?
Moreover, how do the archive access networks impact on the selection
of data compression tools?
- Security (Authentication and Copyright Protection). These topics
are hot ones in the multimedia field, and may apply to remote sensing
data as well. One may want to provide a guarantee to the user, that
the data they received are actually the ones they expected, i.e. nobody
maliciously intercepted the data transmission and modified the data.
Moreover, a data distributor may want to be able to enforce copyright
protection on the distributed data, in order to check that licensing
agreements are not violated by the users through improper use of the
data (e.g., illegal copy and distribution). Which technologies can
be used to this purpose? In the multimedia field digital watermarking
is becoming the key technology, though it arises the same quality
issues as compression. Otherwise, encryption may also be considered.
- Quality assessment. Every time an operation (for example lossy compression)
alters the remote sensing data, e.g. in order to make the data transmission
feasible, a user may be concerned on the quality of the altered data
with respect to the original data or, more in general, on the possible
decrease of added-value. So far, sufficiently general methodologies
to assess data quality do not exist (or they are quite simplistic),
though they are crucial to the deployment of signal processing techniques
on remote sensing data.
- Data Archiving and Distribution will influence standards of relevance
to GRSS.
- This list of research priorities was published in: Archived Imagery: The Undiscovered Country And The Need For Exploration, ESA-EUSC 2004 Workshop on Theory and Applications of Knowledge Driven Image Information Mining, with Focus on Earth Observation, March 17-18, 2004, ESA Special Publication SP-553.
Last updated
on
26-Sep-2004
|
