TIMESAT - A
program to analyse time-series of satellite sensor data.
The TIMESAT pages have been moved. You will be redirected in 5 seconds.
Please update your browser address.
Contents
About TIMESAT
Software development
Authors
Publications
Software distribution policy
Download
File formats
E-mail addresses
About
TIMESAT
TIMESAT was
originally designed to handle noisy time-series of AVHRR NDVI
data and to
extract seasonality information from the data. The program now has the
capability to handle different types of remotely sensed time-series
, e.g. data from Terra/MODIS at different time
resolutions. It has also been tested with eddy covariance data and moisture data, although these
applications are not the main target. Output
from the programme is a set of files containing seasonality (phenology)
parameters
(beginning of season, end of season, amplitude, integrated
values, derivatives, etc.), and smoothed data files.
The function fitting
is done in steps. In the first step the number of seasons and their
approximate timing is defined. The second step filters the data or fits
smooth functions to the data (Savitzky-Golay filter, or least-squares
fitted Gaussian or logistic smooth functions). After the fitting has
been achieved, the seasonality parameters are computed and written to
output files. See the publications
below for more details.
We supply executable for Windows machines and
Fortran source code files. To compile a modified code,
the Intel or Compaq FORTRAN 90 compiler is required. However, the code
is
very close to generic Fortan 90 and can be modified for other compilers
with small effort.
The programme runs from a command prompt under
the operating system using the Fortran executables, or from within
Matlab, versions 6 or 7. We recommend Matlab for tuning of the
parameters and for running small images or time-series. For large
images or long time-series we recommend Fortran.
Software
development
TIMESAT is by no
means a finished product, but is being updated when the need arises,
and when we have available time. We
try to accomodate the needs of individual users and appreciate
suggestions for its improval. TIMESAT includes the
following:
- Core routines in Fortran
and Matlab with processing options for ASCII files and binary images
- A Matlab
graphics interface for simple fitting and parameter settings
- 8-bit integer,
16-bit integer and 32-bit real image formats supported
- Compact internal
data format for phenological data and fitted functions
- Input and fitting
to multiple years of satellite data
- Support for
Gaussian, Savitzky-Golay and logistic functions
- Tools for
generating images from phenological data and fitted functions
Version 2.3 has new and
more stable initial estimation of parameters for the Gaussian and
Logistic fits. All modules have been consistency checked and improved,
and new tools have been added. We have added executables for LINUX.
Authors
Per Jönsson, Malmö University,
Malmö, Sweden
Home page: http://www.mah.se/templates/Page____4464.aspx
Lars Eklundh, Department of Physical Geography and Ecosystems Analysis,
Lund University, Lund, Sweden.
Home page:
http://www.nateko.lu.se/personal/Lars.Eklundh/
Publications
Main publications
describing TIMESAT:
-
Latest software manual: JÖNSSON,
P., and
EKLUNDH, L., 2006, TIMESAT - a Program for Analyzing Time-Series of Satellite Sensor Data. Users Guide for TIMESAT 2.3 (pdf)
- JÖNSSON, P. and EKLUNDH, L., 2004, TIMESAT - a program for analysing time-series of
satellite sensor data, Computers and Geosciences, 30, 833-845. Full
paper (pdf).
- JÖNSSON,
P.
and EKLUNDH, L., 2003, Seasonality extraction from satellite sensor
data. In Frontiers of Remote Sensing Information Processing, edited by
Chen. C.H . World Scientific Publishing. pp 487-500.
- JÖNSSON,
P., and
EKLUNDH, L., 2002, Seasonality extraction by function fitting to
time-series of satellite sensor data. IEEE Transactions on
Geoscience and Remote Sensing, 40 (8), 1824-1832. Full
paper (pdf).
Software
distribution policy
TIMESAT is
distributed for free to the academic community on the
following conditions:
-
The software is the intellectual property of Per Jönsson and Lars Eklundh.
It can be used for free for non-military scientific work by the academic community.
- Users of TIMESAT
in work leading to scientific publications should quote at least our
main
publications (Jönsson and Eklundh, 2002 and Jönsson and
Eklundh, 2004),
see above.
- Co-authorship of
scientific articles should be granted to us in case we extend
substantial help
in setting up, modifying or running TIMESAT.
-
TIMESAT cannot be used for commercial purposes or integrated with commercial software platforms.
- TIMESAT should
not be distributed to others without our consent.
- We should be
informed about the project or reasons for using TIMESAT.
- We do not accept
responsibility for any results from the use of the software.
- We do not
accept liability for any damages due to the use of the software.
- We do not
guarantee user support.
Download
Please check out the Software distribution policy above before proceeding.
To download TIMESAT it is necessary to fill out a registration form. After this form has been completed
the information will be reviewed and a password for downloading will be sent via e-mail, provided that the information is sufficient
and in accordance with our software distribution policy.
To register use the following link: http://timesat.gis.lu.se/
If you have already registered, you can go directly to the login page.
File formats
TIMESAT reads and writes raw, or flat, binary files, one file per
image. Currently 8-bit unsigned integer, 16-bit signed integer or
32-bit signed real data are
supported. Many remote sensing software packages store the data as flat
binary
files with auxilary header files. In most cases these binary files can
be used directly with TIMESAT without any conversion, given that they
are in the correct numeric format (e.g. Idrisi, PCI CIM2, EASI). Data
in software-specific files
that contain header information, are compressed, or organized
differently (BIL, BIP etc.) will have to be converted. File conversion can be carried out with
software such as the open source Geospatial Abstraction Data Base
Library (GDAL), accessed at: http://www.gdal.org/index.html.
lars.eklundh@nateko.lu.se
per.jonsson@mah.se
Page maintenance: Lars Eklundh Dec 07, 2007
