search

astromatic / sextractor

Extract catalogs of sources from astronomical images
http://astromatic.net/software/sextractor
GNU General Public License v3.0
150 stars 44 forks source link

Background subtraction error #14

Open vpicouet opened 4 years ago

vpicouet commented 4 years ago

On the edge of HSC images background subtraction with BACK_SIZE = 64 ends up with only NaNs even though and then detects nothing even though everything works well with 32 or 128… Do you have an idea what causes the problem?

Best


$ sex /Volumes/LaCie/LAM_BABCKUP/calexp/ImagesOnly/calexp-HSC-R-8283-6-8.fits  -PARAMETERS_NAME sex.param  -BACK_SIZE 64
----- SExtractor 2.19.5 started on 2020-05-22 at 12:57:19 with 1 thread

> Setting catalog parameters
> WARNING: Not enough apertures provided in config.:
         some APER photometric values will remain blank 

> Reading Neural Network Weights
> WARNING: Not enough apertures provided in config.:
         some APER photometric values will remain blank 

----- Measuring from: calexp-HSC-R-8283-6-8.fits
      "Unnamed" / no ext. header / 4200x4100 / 32 bits (floats)
(M+D) Background: 0.00500251 RMS: 0.0560535  / Threshold: 0.0840803  
      Objects: detected 0        / sextracted 0               

> All done (in 0.8 s: 5295.8 lines/s , 0.0 detection/s)

$  sex /Volumes/LaCie/LAM_BABCKUP/calexp/ImagesOnly/calexp-HSC-R-8283-6-8.fits  -PARAMETERS_NAME sex.param  -BACK_SIZE 32
----- SExtractor 2.19.5 started on 2020-05-22 at 12:57:24 with 1 thread

> Setting catalog parameters
> WARNING: Not enough apertures provided in config.:
         some APER photometric values will remain blank 

> Reading Neural Network Weights
> WARNING: Not enough apertures provided in config.:
         some APER photometric values will remain blank 

----- Measuring from: calexp-HSC-R-8283-6-8.fits
      "Unnamed" / no ext. header / 4200x4100 / 32 bits (floats)
(M+D) Background: 0.00504417 RMS: 0.0558841  / Threshold: 0.0838262  
      Objects: detected 16799    / sextracted 13263           

> All done (in 2.2 s: 1836.1 lines/s , 5939.4 detections/s)
ebertin commented 4 years ago

The NaNs you are seeing on the background-subtracted check-image? I have tried to reproduce your issue with a 4200×4100 synthetic image but I am not seeing any NaN.

ebertin commented 4 years ago

Same on the image you provided. You might want to check out a more recent version of the software, see this page for getting the most recent version of the code.