IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

source: trunk/psModules/src/imcombine/pmSubtraction.c

Revision Log Mode:


Legend:

Added
Modified
Copied or renamed
Diff Rev Age Author Log Message
(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) @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) @18962   18 years Paul Price Adding threads to pmSubtractionCalculateEquation and …
(edit) @18652   18 years Paul Price More robust estimation of mean, standard deviation.
(edit) @18287   18 years Paul Price Merging in dual-convolution development branch.
(edit) @18190   18 years Paul Price Need to correct for the non-unity scaling of the kernel.
(edit) @18163   18 years Paul Price Changing use of psDB* to pmConfig* so that the database handle only …
(edit) @18043   18 years Paul Price Gene has determined that what I had before --- normalising by the sum …
(edit) @17811   18 years Paul Price Fixing log messages.
(edit) @17783   18 years Paul Price Adding additional analysis metadata.
(edit) @17496   18 years Paul Price Fixing another logical error.
(edit) @17493   18 years Paul Price Fix logical error in image size checking.
(edit) @17394   18 years eugene minor diff
(edit) @17387   18 years Paul Price Fixing logical error.
(edit) @17383   18 years Paul Price More flexibility for inputs to be NULL.
(edit) @17370   18 years Paul Price Making convolution inputs a bit more flexible with mode.
(edit) @17321   18 years Paul Price Fixing API to psImageConvolveFFT.
(edit) @17297   18 years Paul Price Removing MODE_TARGET mode --- this is now done by generating a fake …
(edit) @16607   18 years Paul Price Previous merge didn't have the expected result, probably because of …
(edit) @16604   18 years Paul Price Merging in ppStack development branch --- ppStack now works with …
(edit) @16479   18 years Paul Price Rewinding state of psModules/src/imcombine to the last merge …
(edit) @16476   18 years Paul Price Fixing regions to allow convolution of sub-images (for ppStack)
(edit) @16423   18 years Paul Price Starting to attempt to support stacking of images that have been read …
(edit) @16352   19 years Paul Price I no longer believe that I should be normalising the variance kernel. …
(edit) @15976   19 years gusciora Initialized a couple of pointers to NULL.
(edit) @15809   19 years Paul Price Updates trying to get dual-convolution to work. It doesn't yet.
(edit) @15756   19 years Paul Price Implementing dual-convolution. This required reworking those …
(edit) @15455   19 years Paul Price Don't need second readout if convolving the first.
(edit) @15443   19 years Paul Price Adding function pmSubtractionOrder (choice of name is probably not the …
(edit) @15427   19 years Paul Price Adding additional diagnostics on bad stamps.
(edit) @15422   19 years Paul Price Adding checks on deviations.
(edit) @15390   19 years Paul Price TESTING should be turned off in CVS.
(edit) @15373   19 years Paul Price Adding note about future optimisation.
(edit) @15371   19 years Paul Price Make output coordinates for rejection and deviation consistent.
(edit) @15361   19 years Paul Price Adding more information on stamp rejection to logs.
(edit) @15325   19 years Paul Price Allowing use of FFT convolution for masks. The program was spending a …
(edit) @15319   19 years Paul Price Print the fraction as well.
(edit) @15305   19 years Paul Price Will fail without trying hard if there are not enough good pixels.
(edit) @14870   19 years Paul Price Avoiding error from getting subimage of NULL mask. Plugging memory leak.
(edit) @14867   19 years Paul Price Was getting lots of NANs in unmasked pixels, because the FFT spreads …
(edit) @14824   19 years Paul Price The values in the variance map were about half of what they should …
(edit) @14801   19 years Paul Price Preparing to take a (long) list of sources as candidate stamps. …
(edit) @14753   19 years Paul Price Fixing compilation error.
(edit) @14752   19 years Paul Price Generate kernel for the variance map as the square of the kernel for …
(edit) @14739   19 years Paul Price Give pmSubtractionBorder a size, rather than the NULL kernel (since …
(edit) @14738   19 years Paul Price Consolidating some code. Making it so that the convolutions of the …
(edit) @14716   19 years Paul Price Entire matrix wasn't being populated --- left in some "upper diagonal …
(edit) @14712   19 years Paul Price Fixing compilation error.
(edit) @14711   19 years Paul Price Last commit log message is wrong --- Fixing the matrix generation. …
(edit) @14710   19 years Paul Price Cleaning up, consolidating some duplicated code. Fixing photometric …
(edit) @14709   19 years Paul Price Cleaning up, consolidating some duplicated code. Fixing photometric …
(edit) @14701   19 years Paul Price Extensive changes to image subtraction codes. Discovered that I …
(edit) @14648   19 years magnier insignificant changes
(edit) @14632   19 years Paul Price Moving pmSubtractionDeconvolveMask out of subtraction files, renaming …
(edit) @14624   19 years Paul Price Whoops, removed necessary definition.
(edit) @14623   19 years Paul Price Adding check on return of pmSubtractionKernelImage. Fixing coordinates …
(edit) @14602   19 years Paul Price Adding function pmSubtractionDeconvolveMask that performs matched …
(edit) @14581   19 years Paul Price Fixing normalisation of SPAM,FRIES kernels --- the start/stop values …
(edit) @14540   19 years Paul Price Moving photometric scaling into the kernel for RINGS, and fixing up …
(edit) @14528   19 years Paul Price Ensure correct region is used when cutting down the weight map.
(edit) @14527   19 years Paul Price Use the weight map instead of the input image as the noise term.
(edit) @14524   19 years Paul Price Don't need to give pmSubtractionRejectStamps the images, since those …
(edit) @14515   19 years Paul Price Making the input mask optional for pmSubtractionMask --- necessary if …
(edit) @14513   19 years Paul Price Adding function to set stamps based on a list of x,y positions (as …
(edit) @14480   19 years Paul Price Adding enhancement to extract the pixels for each stamp from the …
(edit) @14456   19 years Paul Price Changing the system by which the kernel weighting is performed. …
(edit) @14455   19 years Paul Price Extensive changes to the way spatial variation of the kernel is …
(edit) @14420   19 years Paul Price Need to normalise sum of kernel components to unity in order to …
(edit) @14360   19 years Paul Price Adding RINGS kernel --- a series of rings, modified by polynomials.
(edit) @14340   19 years Paul Price Should be using the input image as the estimate of the noise for …
(edit) @14329   19 years Paul Price Not using psPolynomials to do spatial polynomials --- it's awkward, …
(edit) @14319   19 years Paul Price CONVOLVE_PIXEL is not needed, since it never needs to do weighting. …
(edit) @14318   19 years Paul Price Fixing MAJOR bug --- the rejection step wasn't doing the convolution …
(edit) @14314   19 years Paul Price Fixing order.
(edit) @14313   19 years Paul Price Putting GUNK in macros.
(edit) @14311   19 years Paul Price Don't need to convolve if the pixel is masked.
(edit) @14305   19 years Paul Price Adding GUNK (Grid United with Normal Kernels --- a POIS/ISIS hybrid) …
(edit) @14304   19 years Paul Price I think I've got it right this time.…
(edit) @14303   19 years Paul Price Fixing constraints on image bounds.
(edit) @14302   19 years Paul Price Adding protection for size of image in convolution. If the kernel …
(edit) @14281   19 years Paul Price Ensure entire image is checked for masked pixels. Can't leave out the …
(edit) @14200   19 years Paul Price Turn off functions in favour of macros.
(edit) @14197   19 years Paul Price Fixing FRIES and SPAM kernels.
(edit) @14195   19 years Paul Price Restoring changes that were wiped when uploading the code to provide …
(edit) @14125   19 years Paul Price Shouldn't use 'x' and 'y' in the macro, since these are defined …
(edit) @14106   19 years Paul Price Adding additional convolution kernels: SPAM (binned pixels around a …
(edit) @13736   19 years Paul Price Fixing range on the mask convolution.
(edit) @13735   19 years Paul Price Unifying input and reference masks in a single mask for the subtraction.
(edit) @13511   19 years Paul Price Checking matrices to ensure no NaNs/Infs creep in. If they do, reject …
(edit) @13406   19 years Paul Price Replacing "middle man" static inline functions with macros, because …
(edit) @13398   19 years Paul Price Optimising the creating of convolution kernels: moved operations that …
(edit) @13395   19 years Paul Price Bit better optimisation on convolution: was generating the kernel …
(edit) @13390   19 years Paul Price Spatial variation working for both POIS and ISIS kernels.
(edit) @13389   19 years Paul Price Fixing subtraction of nominated kernel. This is used to maintain …
(edit) @13386   19 years Paul Price Cleaning up convolvePixel function.
(edit) @13385   19 years Paul Price Removing extraneous line.
(edit) @13380   19 years Paul Price Convolution optimisation, mostly for ISIS kernels.
Note: See TracRevisionLog for help on using the revision log.