Previous release notes

 Site Index
 o Site Home
 o IMG
 o SST


 External Links
 o Durham
 o EAO
 o Starlink

Release note from version 2.3.2

EXTRACTOR has been updated to work with {SExtractor} 2.3.2.

This updates the native SExtractor command to include Large File Support and support for MEFs. The MEF support differs to that previously available in that it now scans a complete MEF file in a single pass.

An error previously limited EXTRACTOR input file names to 20 characters, this has now been corrected.

The EXTRACTOR major version number has been changed to reflect that of the version of SExtractor that it is based on.

Release notes from version 1.4-3

This is a minor release of EXTRACTOR to fix a problems writing out RA and Dec values (this effected EXTRACTOR only, not SExtractor), and a bug in the native version of SExtractor that caused a lockup when writing ASCII_SKYCAT files.

SExtractor has also been upgraded to version 2.2.2 and the license terms of EXTRACTOR have been changed to GPL.

Release notes from version 1.4-0

This is a minor release of EXTRACTOR, just adding two new parameters: X_PIXEL and Y_PIXEL. These output NDF pixel coordinates.

A patch to enable the native version to read MEF files has been incorporated. The extension is identified using a number in square brackets (0 is the primary extension), for instance:

     % $EXTRACTOR_DIR/sex mef_file.fits'[1]'

would process the first extension. MEF files input to SExtractor should not contain a primary image.

Problems and bugs in version 1.4-0

There is a problem running native SExtractor under Redhat 8.0, when writing ASCII_SKYCAT catalogues (the program refuses to exit and goes into a CPU-bound state). Either use another format (like ASCII__HEAD), or upgrade EXTRACTOR.

The news item issued states that MEF files should contain a primary image, rather than not.

Release notes from version 1.3-0

As version 1.2 was never officially released the changes described here actually cover those in 1.2 and 1.3.

Version 1.2-0

This is a minor update to the EXTRACTOR package with several significant changes:

  • SExtractor has been updated to version 2.2.1.
  • New parameters to control the measurement of mean object radii at configurable intensity levels have been added to the Starlink version of EXTRACTOR and SExtractor.

The update to SExtractor 2.2.1 introduces the following changes (from the SExtractor change log).

  • Version 2.2.1
    • Memory allocation bug for large images on Alphas while preparing the background-map fixed.
    • CLEANing bug with hollow objects on weighted images fixed.
  • Version 2.2.0
    • MAP_RMS weight-map bug fixed.
    • Huge MAP_WEIGHT calibration bugs fixed.
  • Version 2.1.6
    • Crowding-flag bug introduced in V2.1.5 fixed.
  • Version 2.1.5
    • Small inaccuracy in crowding-flag positioning fixed.
    • Fixed division by zero in local background in MASK_TYPE CORRECT mode.
    • Risk of segmentation fault with huge weight maps fixed.
  • Version 2.1.4
    • Quote (') symbol now properly handled in FITS headers.
    • Removed display of debug information inadvertently left in the previous release.
  • Version V2.1.3
    • Bug in ASSOC on Linux systems fixed.
    • Background bug in WEIGHT_TYPE NONE, MAP_WEIGHT mode fixed.
    • THRESH_TYPE ABSOLUTE now works on weighted images.
  • Version V2.1.0
    • Better formatting of output flags.
    • New display.
    • Total robust handling of weight-maps during background- determination.
    • (tiny) differences between single and double-image results suppressed.
    • New (experimental) _PROFILE photometric parameters.
    • New WEIGHT_GAIN parameter.
  • Version V2.0.22
    • Many memory leaks fixed.
    • CLEANing bug with very faint sources fixed.
    • BACKGROUND_RMS check-images are now operational in single-image mode.
    • Documentation improved.
  • Version V2.0.21
    • BACK_VALUE bug fixed.
    • XPEAK and YPEAK offset bugs fixed.
  • Version V2.0.20
    • ASSOC bug in 2 ASSOC_PARAMS mode fixed.
    • Variable thresholding is now taken into account when computing FWHM and CLASS_STAR if WEIGHT_TYPE != NONE

Version 1.3-0

A minor update to EXTRACTOR with the following changes:

  • a bug in EXTRACTOR that could cause a crash on Tru64 UNIX if the measurement image contained BAD values that the detection image did not, has been fixed.
  • The output format for the X_WORLD and Y_WORLD fields has been changed to always give at least 0.01 arcsecond accuracy.
  • EXTRACTOR and SExtractor have modified to offer the BKGSIG parameter (local background sigma) option used by GIM2D .
  • A bug causing the output from EXTRACTOR to be occasionally garbled has been fixed.

Release notes from version 1.1-3

  • The range of output WCS coordinates should now always be correctly normalized (i.e. RA's in range 0-24hrs).
  • Apertures should now be actually visible in the "APERTURES" checkimage.
  • Skycat catalogues output from detached monoliths (Tcl & ICL) should now always have the correct Tab formatting. This caused the "blank catalogue window" problems in GAIA when repeatedly analyzing the same image.

Problems and bug from 1.1-3

There is a problem with EXTRACTOR on Tru64 UNIX if the measurement image contains BAD values that the detection image did not.

There is a bug that causes the output from EXTRACTOR to be garbled which can then be followed by a crash (only reported for Linux).

The only work arounds for these was to upgrade.

Release notes from version 1.1-1

The EXTRACTOR and SExtractor programs have been updated to SExtractor version 2.0.19 (the previous version was 2.0.15). From the SExtractor change log:

  • SEGMENTATION numbering problem fixed
  • new ASSOC_TYPE NEAREST option.
  • MAP_RMS measurement mode and FLAG_TYPE MAX bugs fixed.
  • new BACK_TYPE, BACK_VALUE and THRESH_TYPE parameters.
  • FLUX_RADIUS no longer requires FLUX_AUTO.
  • Documentation (mud165) slightly improved.

The BACK_TYPE/BACK_VALUE parameters now allow the background used to be defined as a single constant for the whole image (BACK_TYPE takes the values "INTERNAL" and "MANUAL") and the THRESH_TYPE controls whether the detection and analysis thresholds are a multiple of the background RMS, or just in absolute data units.

A serious bug calculating the CLASS_STAR parameter (which is only revealed on Digital UNIX systems) has been fixed.

The MAG_APER parameter now works correctly under ICL (and TCL). This was crashing during a second run.

Problems and bugs from 1.1-1

  • The world coordinates output by EXTRACTOR may not be normalized correctly (i.e. you may see negative RAs).
  • The "APERTURES" checkimage is broken.
  • Repeated object detection scans in GAIA (and ICL) using a SKYCAT output catalogue fail to add the correct formatting to the data columns. This results in a failure to see the catalogue on the second pass.

There are no work arounds for these bugs. Obtain a new release.

Questions or comments to:

Copyright © 2005 Central Laboratory of the Research Councils
Copyright © 2006 Particle Physics and Astronomy Research Council
Copyright © 2008-2009 Science and Technlogy Facilities Council
Copyright © 2009-2013 Peter W. Draper
Last modified: 02-Jun-2016
This site generated by htmlPX.