IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

source: trunk/psModules/src/imcombine

Revision Log Mode:


Legend:

Added
Modified
Copied or renamed
Diff Rev Age Author Log Message
(edit) @21469   17 years Paul Price Transfer covariance factor to variance image.
(edit) @21429   17 years Paul Price Fixing type name to be compatible with macro use.
(edit) @21428   17 years Paul Price Fixing kernel for variance: not supposed to include the (sum k_ij)2 …
(edit) @21422   17 years Paul Price Merging cnb_branch_20090113. No major conflicts. Compiles, but not …
(edit) @21363   17 years Paul Price Merging pap_branch_20090128. Resolved a small number of conflicts. …
(edit) @21351   17 years eugene comment out the mutex spins on the psImages (now handled by the …
(edit) @21247   17 years eugene fix vector mask type; disable apparent test code (forced return of …
(edit) @21183   17 years eugene incorporating changes from 16bit mask upgrades (eam_branch_20081230)
(edit) @21149   17 years Paul Price Adding test for deconvolution: measures the fraction of the kernel …
(edit) @20999   18 years Paul Price Attempting to get correct scaling for stacked images.
(edit) @20937   18 years eugene changing pmPeaks to contain a pointer to pmFootprints, which just …
(edit) @20838   18 years Paul Price Ensure thread jobs have the expected type when harvesting.
(edit) @20832   18 years Paul Price Turn psImageConvolveMask threads on/off as required, and set back when …
(edit) @20718   18 years Paul Price Fixing "may be used uninitialised" error.
(edit) @20717   18 years Paul Price Fixing double free.
(edit) @20714   18 years Paul Price Apply threshold to stamps from list. Threshold is now specified as …
(edit) @20712   18 years eugene var_soft = var + soft_frac nsigma_2 = var_soft * rej (not the other …
(edit) @20633   18 years Paul Price Adding additional TESTING outputs.
(edit) @20571   18 years Paul Price NAN sysError was causing trouble.
(edit) @20568   18 years Paul Price Adding parameter 'stride' to control the size of convolution patches …
(edit) @20561   18 years Paul Price Adding log messages with time of important stages.
(edit) @20505   18 years Paul Price Making relative systematic error in kernel a configurable parameter …
(edit) @20497   18 years Paul Price Adding relative systematic error for combination when rejecting.
(edit) @20487   18 years Paul Price Don't use variance factors since we're now renormalising the weight maps.
(edit) @20466   18 years Paul Price Need to turn off threads at the psLib level --- otherwise, we end up …
(edit) @20465   18 years Paul Price Don't rely on the input source list to have finite magnitudes (at the …
(edit) @20421   18 years Paul Price Fixing SEGV when tracing due to optimised ISIS.
(edit) @20420   18 years Paul Price Adding trace statements. Bug resulted in zero pixels rejected when …
(edit) @20355   18 years Paul Price Write out realisations of the kernel as well.
(edit) @20215   18 years Paul Price Turning off systematic error --- doesn't have a large effect.
(edit) @20052   18 years Paul Price Was working with an image, not a kernel.
(edit) @20051   18 years Paul Price Fixing typo.
(edit) @20050   18 years Paul Price Adding quality of fit information.
(edit) @20049   18 years Paul Price Adding new file with function to add stuff to a common 'analysis' …
(edit) @19998   18 years eugene test for valid model
(edit) @19794   18 years Paul Price More consistent about which solution goes where.
(edit) @19768   18 years Paul Price Removing old code.
(edit) @19767   18 years Paul Price Restoring calculation of moments for ISIS.
(edit) @19766   18 years Paul Price Free systematic error images.
(edit) @19765   18 years Paul Price Two major changes. 1. Optimising ISIS kernels: instead of doing an …
(edit) @19758   18 years Paul Price Always use the same name.
(edit) @19669   18 years Paul Price Removing MEAN, RMS, NUM --- they go in the kernel now, not the …
(edit) @19541   18 years Paul Price Making use of optimised psPixelsConcatenate.
(edit) @19540   18 years Paul Price Dereferencing in the hope things will go faster.
(edit) @19535   18 years Paul Price Writing a number as float instead of int.
(edit) @19532   18 years Paul Price Correcting stack variances for the imperfect convolution. Changed …
(edit) @19487   18 years Paul Price Don't use variance factors.
(edit) @19482   18 years Paul Price Merging in pap_branch_080908. Was originally going to use it for …
(edit) @19357   18 years Paul Price Always record the RMS.
(edit) @19349   18 years Paul Price Fixed bug: wasn't initialising target image.
(edit) @19345   18 years Paul Price Threading pmStackReject
(edit) @19344   18 years Paul Price Can only lock if pointer is non-NULL.
(edit) @19340   18 years Paul Price Don't need mutexes on output images.
(edit) @19336   18 years Paul Price Create empty pixel list before starting combination.
(edit) @19335   18 years Paul Price Moving propagation of mask into convolveRegion(): it's already going …
(edit) @19299   18 years Paul Price psThreadJob is now mutable within the thread function.
(edit) @19282   18 years Paul Price We want a bad/poor mask distinction in the mask convolution for …
(edit) @19271   18 years Paul Price Reducing (and asserting on) bounds for copy. There's some memory …
(edit) @19247   18 years Paul Price Remove extraneous macro definitions.
(edit) @19243   18 years Paul Price Trying to reduce memory usage, which gets somewhat large in the end of …
(edit) @19241   18 years Paul Price Don't increment pixel array size bit by bit.
(edit) @19239   18 years Paul Price Caught an important bug: was using the wrong index in comparison.
(edit) @19234   18 years Paul Price Revising source mask.
(edit) @19230   18 years Paul Price In the case that there is no chip/cell (the readout is orphaned), the …
(edit) @19225   18 years Paul Price Adding additional debugging output when TESTING.
(edit) @19209   18 years Paul Price 'inner' member of pmSubtractionKernels wasn't being updated.
(edit) @19207   18 years Paul Price Avoiding use of sqrt.
(edit) @19205   18 years Paul Price Just in case…
(edit) @19204   18 years Paul Price Fixing compilation error. Didn't set sizes.
(edit) @19202   18 years Paul Price More restrictions on sources.
(edit) @19201   18 years Paul Price Don't blindly use all input sources as stamps.
(edit) @19189   18 years Paul Price Adding I/O of subtraction kernels to pmFPAfile stuff.
(edit) @19170   18 years Paul Price Reject anomalously low deviations as well.
(edit) @19169   18 years Paul Price Turn off testing for CVS copy!
(edit) @19168   18 years Paul Price Need more iterations!
(edit) @19164   18 years Paul Price Reworking the masking so as to avoid exploding the number of bad …
(edit) @19148   18 years Paul Price Propagate values from inputs masks.
(edit) @19122   18 years Paul Price Need to put mutex locks on the parent images when convolving, since …
(edit) @19086   18 years eugene test for valid image on each readout
(edit) @19061   18 years Paul Price Only lock if running threaded.
(edit) @19059   18 years Paul Price Adding threading for pmSubtractionConvolve. This was more difficult …
(edit) @19009   18 years Paul Price Another installment in the saga to ensure pmSourceMoments doesn't go …
(edit) @18962   18 years Paul Price Adding threads to pmSubtractionCalculateEquation and …
(edit) @18861   18 years eugene comment out timer marks
(edit) @18830   18 years eugene src/objects/Makefile.am
(edit) @18788   18 years Paul Price Work out maximum "valid" radius (for S/N ~ 1; which is what's used by …
(edit) @18652   18 years Paul Price More robust estimation of mean, standard deviation.
(edit) @18611   18 years Paul Price Need to use Poisson errors in order to get realistic weight values.
(edit) @18566   18 years Paul Price Resolving conflicts.
(edit) @18508   18 years Paul Price Fix to allow an image to have no pixels to inspect/reject.
(edit) @18492   18 years Paul Price Trying to get a better system for setting the weightings.
(edit) @18348   18 years Paul Price Positions are supposed to be normalised (-1,1).
(edit) @18347   18 years Paul Price Don't blow away the kernels before we need them!
(edit) @18287   18 years Paul Price Merging in dual-convolution development branch.
(edit) @18202   18 years Paul Price region may be NULL.
(edit) @18201   18 years Paul Price region may be NULL.
(edit) @18190   18 years Paul Price Need to correct for the non-unity scaling of the kernel.
(edit) @18182   18 years Paul Price Incorporating "variance factors" --- accounting for pixel-scale …
(edit) @18181   18 years Paul Price Record the variance factor.
(edit) @18180   18 years Paul Price Adding pmSubtractionVarianceFactor.
Note: See TracRevisionLog for help on using the revision log.