|
|
|
@24622
|
17 years |
eugene |
use 1 sigma clipping on source moments (probably not needed)
|
|
|
|
@24621
|
17 years |
eugene |
added some test code
|
|
|
|
@24620
|
17 years |
eugene |
added some test code
|
|
|
|
@24591
|
17 years |
eugene |
update to pmSourceMoments API -- turn off S/N clipping
|
|
|
|
@24298
|
17 years |
Paul Price |
Fixing application of ISIS and POIS kernels to stamps: was effectively …
|
|
|
|
@24297
|
17 years |
Paul Price |
Adding note about threads when TESTING.
|
|
|
|
@24296
|
17 years |
Paul Price |
Fixing moment following fix to sum.
|
|
|
|
@24295
|
17 years |
Paul Price |
Fixing normalisation: was normalising both x and y by the sum of the …
|
|
|
|
@24293
|
17 years |
Paul Price |
Include variance in matrix equation.
|
|
|
|
@24292
|
17 years |
Paul Price |
Allow zero iterations.
|
|
|
|
@24259
|
17 years |
Paul Price |
Use last code instead of default.
|
|
|
|
@24257
|
17 years |
Paul Price |
Adding new error code to treat bad data quality: SMALL_AREA = …
|
|
|
|
@24212
|
17 years |
bills |
Turn off TESTING
|
|
|
|
@24206
|
17 years |
Paul Price |
Fitting a trend of chi2 as a function of flux was often failing when …
|
|
|
|
@24181
|
17 years |
Paul Price |
Fixing ticket 1246: inverting preference order for stamps source …
|
|
|
|
@24093
|
17 years |
eugene |
rename LU decomp functions
|
|
|
|
@24066
|
17 years |
Paul Price |
Search around provided stamp centre for a suitable stamp. If there's …
|
|
|
|
@23989
|
17 years |
eugene |
always handle psVectorStats false return status (is an error); check …
|
|
|
|
@23975
|
17 years |
Paul Price |
Turn off TESTING for repository version... again.
|
|
|
|
@23973
|
17 years |
Paul Price |
Didn't undo all of the maxRadius changes yesterday.
|
|
|
|
@23961
|
17 years |
Paul Price |
We do need the source radii, otherwise the moments all fail.
|
|
|
|
@23960
|
17 years |
Paul Price |
Doing our best to cope when working with skycells on the edge of the …
|
|
|
|
@23956
|
17 years |
Paul Price |
Turn off TESTING for repository.
|
|
|
|
@23955
|
17 years |
Paul Price |
Catching bad source models (using flags) and ejecting these from the …
|
|
|
|
@23944
|
17 years |
Paul Price |
Fix double free.
|
|
|
|
@23937
|
17 years |
Paul Price |
Implement stamp thresholds on both images. This is to avoid …
|
|
|
|
@23851
|
17 years |
Paul Price |
Convolve subtraction mask so that we can easily identify bad stamps.
|
|
|
|
@23850
|
17 years |
Paul Price |
If stamp point is outside border, then don't need to mark it as bad.
|
|
|
|
@23849
|
17 years |
Paul Price |
Fixing how stamps are found: ensuring NANs don't get into the stamp …
|
|
|
|
@23839
|
17 years |
Paul Price |
Generate ds9 region files for stamps at interesting places.
|
|
|
|
@23837
|
17 years |
Paul Price |
Trigger region output based on trace level.
|
|
|
|
@23792
|
17 years |
Paul Price |
Fixing warning.
|
|
|
|
@23791
|
17 years |
Paul Price |
Fixing warning regarding uninitialised use.
|
|
|
|
@23789
|
17 years |
Paul Price |
Adding first cut at function to write ds9 region file with stamp positions.
|
|
|
|
@23780
|
17 years |
Paul Price |
For deconvolution fraction, record only the maximum. By putting it in …
|
|
|
|
@23775
|
17 years |
Paul Price |
Turn off debugging for SVN.
|
|
|
|
@23773
|
17 years |
Paul Price |
Weighted Olympic mean wasn't working as intended: high pixels weren't …
|
|
|
|
@23740
|
17 years |
Paul Price |
Merging branches/pap from r23739 onto trunk. No conflicts, ppSub …
|
|
|
|
@23688
|
17 years |
Paul Price |
Merging in branches/pap/ from r23685. Some conflicts, most notably in …
|
|
|
|
@23577
|
17 years |
Paul Price |
Ensure we can reject pixels when combining images without convolution.
|
|
|
|
@23487
|
17 years |
Paul Price |
Functions declared as 'type function()' cause errors with the Intel …
|
|
|
|
@23378
|
17 years |
Paul Price |
Removing unnecessary restriction that is preventing kernel files being …
|
|
|
|
@23374
|
17 years |
Paul Price |
Put top-level HDU (should be PHU) into header, so that output can be …
|
|
|
|
@23308
|
17 years |
Paul Price |
It's a little more complicated than I thought --- there could be …
|
|
|
|
@23307
|
17 years |
Paul Price |
Adding function pmSubtractionMatchPrecalc to convolve using a kernel …
|
|
|
|
@23259
|
17 years |
Paul Price |
Changing API for psRandomAlloc() to make it easier to be …
|
|
|
|
@23242
|
17 years |
beaumont |
merging from cnb_branch_20090215
|
|
|
|
@23191
|
17 years |
Paul Price |
Allow input PSFs to be NULL.
|
|
|
|
@22715
|
17 years |
Paul Price |
Removing unnecessary .cvsignore files.
|
|
|
|
@22699
|
17 years |
Paul Price |
Reorganised the concepts files, moving the Write functions into …
|
|
|
|
@21476
|
17 years |
Paul Price |
Using covariance factor to set the correct variance level. Median …
|
|
|
|
@21470
|
17 years |
Paul Price |
Actually, don't transfer. Let the user do it, since he may (does, in …
|
|
|
|
@21469
|
17 years |
Paul Price |
Transfer covariance factor to variance image.
|
|
|
|
@21429
|
17 years |
Paul Price |
Fixing type name to be compatible with macro use.
|
|
|
|
@21428
|
17 years |
Paul Price |
Fixing kernel for variance: not supposed to include the (sum k_ij)2 …
|
|
|
|
@21422
|
17 years |
Paul Price |
Merging cnb_branch_20090113. No major conflicts. Compiles, but not …
|
|
|
|
@21363
|
17 years |
Paul Price |
Merging pap_branch_20090128. Resolved a small number of conflicts. …
|
|
|
|
@21351
|
17 years |
eugene |
comment out the mutex spins on the psImages (now handled by the …
|
|
|
|
@21247
|
17 years |
eugene |
fix vector mask type; disable apparent test code (forced return of …
|
|
|
|
@21183
|
17 years |
eugene |
incorporating changes from 16bit mask upgrades (eam_branch_20081230)
|
|
|
|
@21149
|
17 years |
Paul Price |
Adding test for deconvolution: measures the fraction of the kernel …
|
|
|
|
@20999
|
18 years |
Paul Price |
Attempting to get correct scaling for stacked images.
|
|
|
|
@20937
|
18 years |
eugene |
changing pmPeaks to contain a pointer to pmFootprints, which just …
|
|
|
|
@20838
|
18 years |
Paul Price |
Ensure thread jobs have the expected type when harvesting.
|
|
|
|
@20832
|
18 years |
Paul Price |
Turn psImageConvolveMask threads on/off as required, and set back when …
|
|
|
|
@20718
|
18 years |
Paul Price |
Fixing "may be used uninitialised" error.
|
|
|
|
@20717
|
18 years |
Paul Price |
Fixing double free.
|
|
|
|
@20714
|
18 years |
Paul Price |
Apply threshold to stamps from list. Threshold is now specified as …
|
|
|
|
@20712
|
18 years |
eugene |
var_soft = var + soft_frac
nsigma_2 = var_soft * rej
(not the other …
|
|
|
|
@20633
|
18 years |
Paul Price |
Adding additional TESTING outputs.
|
|
|
|
@20571
|
18 years |
Paul Price |
NAN sysError was causing trouble.
|
|
|
|
@20568
|
18 years |
Paul Price |
Adding parameter 'stride' to control the size of convolution patches …
|
|
|
|
@20561
|
18 years |
Paul Price |
Adding log messages with time of important stages.
|
|
|
|
@20505
|
18 years |
Paul Price |
Making relative systematic error in kernel a configurable parameter …
|
|
|
|
@20497
|
18 years |
Paul Price |
Adding relative systematic error for combination when rejecting.
|
|
|
|
@20487
|
18 years |
Paul Price |
Don't use variance factors since we're now renormalising the weight maps.
|
|
|
|
@20466
|
18 years |
Paul Price |
Need to turn off threads at the psLib level --- otherwise, we end up …
|
|
|
|
@20465
|
18 years |
Paul Price |
Don't rely on the input source list to have finite magnitudes (at the …
|
|
|
|
@20421
|
18 years |
Paul Price |
Fixing SEGV when tracing due to optimised ISIS.
|
|
|
|
@20420
|
18 years |
Paul Price |
Adding trace statements. Bug resulted in zero pixels rejected when …
|
|
|
|
@20355
|
18 years |
Paul Price |
Write out realisations of the kernel as well.
|
|
|
|
@20215
|
18 years |
Paul Price |
Turning off systematic error --- doesn't have a large effect.
|
|
|
|
@20052
|
18 years |
Paul Price |
Was working with an image, not a kernel.
|
|
|
|
@20051
|
18 years |
Paul Price |
Fixing typo.
|
|
|
|
@20050
|
18 years |
Paul Price |
Adding quality of fit information.
|
|
|
|
@20049
|
18 years |
Paul Price |
Adding new file with function to add stuff to a common 'analysis' …
|
|
|
|
@19998
|
18 years |
eugene |
test for valid model
|
|
|
|
@19794
|
18 years |
Paul Price |
More consistent about which solution goes where.
|
|
|
|
@19768
|
18 years |
Paul Price |
Removing old code.
|
|
|
|
@19767
|
18 years |
Paul Price |
Restoring calculation of moments for ISIS.
|
|
|
|
@19766
|
18 years |
Paul Price |
Free systematic error images.
|
|
|
|
@19765
|
18 years |
Paul Price |
Two major changes. 1. Optimising ISIS kernels: instead of doing an …
|
|
|
|
@19758
|
18 years |
Paul Price |
Always use the same name.
|
|
|
|
@19669
|
18 years |
Paul Price |
Removing MEAN, RMS, NUM --- they go in the kernel now, not the …
|
|
|
|
@19541
|
18 years |
Paul Price |
Making use of optimised psPixelsConcatenate.
|
|
|
|
@19540
|
18 years |
Paul Price |
Dereferencing in the hope things will go faster.
|
|
|
|
@19535
|
18 years |
Paul Price |
Writing a number as float instead of int.
|
|
|
|
@19532
|
18 years |
Paul Price |
Correcting stack variances for the imperfect convolution. Changed …
|
|
|
|
@19487
|
18 years |
Paul Price |
Don't use variance factors.
|
|
|
|
@19482
|
18 years |
Paul Price |
Merging in pap_branch_080908. Was originally going to use it for …
|
|
|
|