6.5.11    APEX Modules: Bright Detect

Command Line Equivalent: run_bright_detect

Default Output Directory: <output_dir>; <output_dir>/Coadd-apex

Depends On: Initial Setup

Relevant Pipelines: APEX Multiframe; APEX Single Frame

 

PURPOSE

This module provides a different kind of detection table than the Detect. It includes shape information about the detections, but it does not record the blends, so in general it cannot replace Detect.  It also produces a mask image (pixel value = 1 on the detections).  This can be used to mask the brightest detections before running Detect.

 

INPUT

Same as for the Detect module (see next section, §6.5.12).

 

COMMAND LINE INPUT

See the Detect module

 

OUTPUT

Table (*_detect_bright.tbl): a detection list with shape information.

 

FITS file (*_detect_bright_mask.fits): a mask image showing the segmentation level of detected clusters.

 

For APEX Multiframe, the output files are written to the Coadd-apex/ subdirectory.

 

 

DISCUSSION

The bright detect list created by this module provides information about the shapes of the detections. The shape parameters are mostly clear from the column names. The shape parameter Linearity is 0.25 * (Perimeter length) / sqrt (No. pixels). Below is an example of the detection table:

 

\char comment = Output from DETECT, version 4.00

\char Date-Time = Wed Feb 4 12:12:34 2004

\char comment = dlimage library version 2.00

\char FITS_In_Filename = Coadd-apex/coadd_Tile_001_Image_minback.fits

\char FITS_Out_Filename = Coadd-apex/coadd_Tile_001_Image_detect_bright_mask.fits

\char Sigma_In_Filename = Coadd-apex/coadd_Tile_001_Unc.fits

\char Bright_Table_Out_Filename = Coadd-apex/coadd_Tile_001_Image_detect_bright.tbl

\float Tile_OffsetX = 0

\float Tile_OffsetY = 0

\int Peaks_Radius = 1

\char CoverageMap_Filename = Coadd-apex/coadd_Tile_001_Cov.fits

\float Min_Coverage = 2

\float Effective_Threshold = 0.259142

\int Extended_Object_Area = 200

\int Max_Segmentation_Level = 50

\int Detection_Max_Area = 1000

\int Detection_Min_Area = 30

\int N_Edge = 0

\float Detection_Threshold = 6

\char Input_Type = image_input

\char Output_Type = centroids_and_pixels_output

\char Neighbor_Type = sides_only

\char Threshold_Type = peak

\int Number_Of_Detections = 118

srcid| x       | delta_x | y       | delta_y  | delta_xy | flux    | delta_flux| maj_ax  | delta_maj| min_ax  | delta_min| elongation| ellipticity| theta    | delta_theta| n_pixels| perimeter| linearity|

|  i |r        |      r  |r        |      r   |       r  |   r     |         r |     r   |        r |     r   |        r |         r |          r |    r     |          r |       i |        i |        r |

  1   6.292e+02 6.751e-02 2.552e+01 7.353e-02  3.333e-02  8.223e+02 1.407e+01   3.118e+00 7.856e -02 2.515e+00 6.159e-02  1.240e+00   1.934e-01    5.636e+01  5.636e+01    332        94        1.290e+00

  2   1.493e+02 1.350e-02 3.395e+01 1.430e-02  -4.271e-03 3.814e+03 2.266e+01   1.553e+00 1.465e -02 1.471e+00 1.312e-02  1.056e+00   5.289e-02    -6.515e+01 -6.063e+01   278        132       1.979e+00

  3   1.232e+03 3.253e-02 3.493e+01 3.006e-02  1.101e-03  3.814e+02 7.638e+00   1.118e+00 3.354e -02 1.020e+00 2.893e-02  1.095e+00   8.712e-02    1.720e+01  4.057e+01     70         46       1.375e+00

 

The bright detect mask marks detections with integer values equal to the highest segmentation level (iteration) reached, and non-detection pixels with 0.

 

Bright Detect can provide useful information as an adjunct to the normal Detect module. In this case, set the available parameters to the same as those used for Detect, but be aware that because it does not include blend parameters, it cannot replace Detect for source extraction.

 

Advanced users could use this to try to mask out the brightest sources before running Detect. Set the Detection Threshold to a high value and turn on Use Bright Object Mask in the APEX Settings module. This tells Detect to mask out the bright detections. Note: This feature has not been well tested and users should exercise caution. 

• 
• 
• 
• 
•