[LSST|dm-users 644] Re: Detection and Measurement of Postage Stamp
Russell Owen
rowen at uw.edu
Mon May 11 16:28:53 PDT 2015
Yes there are a lot of afw packages, but I’m hoping you find you can focus on a few. The doxygen documentation is a good place to find out where a particular tasks lives (especially the page that lists each task). Many tasks are in pipe_tasks and meas_algorithms. A few tasks are in ip_diffim, ip_isr and other places. Camera-specific tasks are in obs_* packages.
Regards,
— Russell
On May 11, 2015, at 4:24 PM, Will Dawson <will at dawsonresearch.com> wrote:
> Russell,
>
> I am comfortable with github but I am not entirely comfortable navigating the 10 pages of repositories in https://github.com/lsst, and glancing at lsst/afw isn’t comforting. With some guidance on which repo(s) to operate in this may be a workable solution though.
>
> Thanks,
> Will
>
>> On May 11, 2015, at 4:11 PM, Russell Owen <rowen at uw.edu> wrote:
>>
>> I think it would be great if you were willing to improve the documentation. The simplest approach, if you are at all comfortable with github, is probably for you to fork the code on github and commit fixes to your fork, then submit a pull request. If not, then I imagine we can think of something simpler for you.
>>
>> — Russell
>>
>> On May 11, 2015, at 3:50 PM, Will Dawson <will at dawsonresearch.com> wrote:
>>
>>> Jim / Robert,
>>>
>>> I’ve been running everything from python so I guess I will stick to help(). Is there a way that I can help edit the documentation as I learn things? For example I could help add some meat to things like:
>>>
>>> reEstimateBackground
>>> | Estimate the background again after final source detection?
>>>
>>> Although I completely understand if you don’t want people who are just learning things to be editing the help documentation.
>>>
>>> Thanks again,
>>> Will
>>>
>>>> On May 11, 2015, at 3:13 PM, Jim Bosch <jbosch at astro.princeton.edu> wrote:
>>>>
>>>> The Doxygen is good for some things, but the configuration of Python code isn't one of them. If you're running a command-line task, then Robert's suggestion is the way to go. If not, I recommend just using Python's help() function on the config object - that should get you the list of options and some documentation for each of them.
>>>>
>>>> Jim
>>>>
>>>>
>>>> On Mon, May 11, 2015 at 6:10 PM, Robert Lupton the Good <rhl at astro.princeton.edu> wrote:
>>>> I don't know how you are running your task, but if it's using commandLineTask then
>>>> --show=*[Bb]ackground*
>>>> should show all configuration parameters matching the glob. (Benighted [t]csh users will need to quote the pattern)
>>>>
>>>> R
>>>>
>>>> On 11 May 2015, at 18:00, Jim Bosch <jbosch at astro.princeton.edu> wrote:
>>>>
>>>>> I think this should be just a matter of setting
>>>>>
>>>>> detectSourcesConfig.reEstimateBackground = False
>>>>>
>>>>>
>>>>>
>>>>> Jim
>>>>>
>>>>>
>>>>>
>>>>> On Mon, May 11, 2015 at 5:41 PM, Will Dawson <will at dawsonresearch.com> wrote:
>>>>> I am trying to use DM Stack detect and analyze in object in a small postage stamp (see attached example, log scale). I know I can use lsst.afw.detection.Threshold() with lsst.afw.detection.FootprintSet() to create a footprint for the detected object. But I also (first) tried lsst.meas.algorithms.SourceDetectionTask(config=detectSourcesConfig, "other args") with config set as:
>>>>>
>>>>> detectSourcesConfig = measAlg.SourceDetectionConfig(thresholdType='value’)
>>>>> or
>>>>> detectSourcesConfig = measAlg.SourceDetectionConfig(thresholdType='value',thresholdValue=0)
>>>>>
>>>>> and got the error at the end of this email. I am not entirely sure how to parse this error statement, but if I had to guess: no objects are being detected because it is trying to estimate the background based on this stamp which contains primary just the object thus when it does image - background there is nothing above threshold left to detect. Or it chokes when trying to estimate the background.
>>>>>
>>>>> Is there an option to run lsst.meas.algorithms.SourceDetectionTask with the background estimation/ subtraction process turned off? For example when detecting objects on an image that already has the background subtracted. Perhaps the only option is to stick with the detection.FootprintSet method.
>>>>>
>>>>> Thanks,
>>>>> Will
>>>>>
>>>>>
>>>>> <28x28stamp.png>
>>>>>
>>>>> ---------------------------------------------------------------------------
>>>>> InvalidParameterError Traceback (most recent call last)
>>>>> <ipython-input-66-d08d86b9540f> in <module>()
>>>>> 2 # catalog atually stores stuff)
>>>>> 3 table = afwTable.SourceTable.make(schema)
>>>>> ----> 4 catalog = detect.makeSourceCatalog(table, exposure, sigma=3)
>>>>>
>>>>> /Users/dawson/lsst/DarwinX86/pipe_base/10.0+2/python/lsst/pipe/base/timer.pyc in wrapper(self, *args, **keyArgs)
>>>>> 116 logInfo(obj = self, prefix = func.__name__ + "Start")
>>>>> 117 try:
>>>>> --> 118 res = func(self, *args, **keyArgs)
>>>>> 119 finally:
>>>>> 120 logInfo(obj = self, prefix = func.__name__ + "End")
>>>>>
>>>>> /Users/dawson/lsst/DarwinX86/meas_algorithms/10.0+2/python/lsst/meas/algorithms/detection.pyc in run(self, table, exposure, doSmooth, sigma, clearMask)
>>>>> 313 raise ValueError("Table has incorrect Schema")
>>>>> 314 fpSets = self.detectFootprints(exposure=exposure, doSmooth=doSmooth, sigma=sigma,
>>>>> --> 315 clearMask=clearMask)
>>>>> 316 sources = afwTable.SourceCatalog(table)
>>>>> 317 table.preallocate(fpSets.numPos + fpSets.numNeg) # not required, but nice
>>>>>
>>>>> /Users/dawson/lsst/DarwinX86/pipe_base/10.0+2/python/lsst/pipe/base/timer.pyc in wrapper(self, *args, **keyArgs)
>>>>> 116 logInfo(obj = self, prefix = func.__name__ + "Start")
>>>>> 117 try:
>>>>> --> 118 res = func(self, *args, **keyArgs)
>>>>> 119 finally:
>>>>> 120 logInfo(obj = self, prefix = func.__name__ + "End")
>>>>>
>>>>> /Users/dawson/lsst/DarwinX86/meas_algorithms/10.0+2/python/lsst/meas/algorithms/detection.pyc in detectFootprints(self, exposure, doSmooth, sigma, clearMask)
>>>>> 443 fpSets.background = bkgd
>>>>> 444 self.log.log(self.log.INFO, "Resubtracting the background after object detection")
>>>>> --> 445 mi -= bkgd.getImageF()
>>>>> 446 del mi
>>>>> 447
>>>>>
>>>>> /Users/dawson/lsst/DarwinX86/afw/10.0+3/python/lsst/afw/math/mathLib.pyc in getImageF(self, *args)
>>>>> 5736 getImageF(Background self) -> boost::shared_ptr< lsst::afw::image::Image< float > >
>>>>> 5737 """
>>>>> -> 5738 return _mathLib.Background_getImageF(self, *args)
>>>>> 5739
>>>>> 5740 def getImageI(self, *args):
>>>>>
>>>>> InvalidParameterError:
>>>>> File "src/math/Interpolate.cc", line 60, in std::pair<std::vector<double>, std::vector<double> > lsst::afw::math::(anonymous namespace)::recenter(const std::vector<double> &, const std::vector<double> &)
>>>>> You must provide at least 1 point {0}
>>>>> File "src/math/BackgroundMI.cc", line 196, in void lsst::afw::math::BackgroundMI::_setGridColumns(const Interpolate::Style, const lsst::afw::math::UndersampleStyle, const int, const std::vector<int> &) const
>>>>> setting _gridcolumns {1}
>>>>> lsst::pex::exceptions::InvalidParameterError: 'You must provide at least 1 point {0}; setting _gridcolumns {1}'
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> dm-users mailing list
>>>>> dm-users at lists.lsst.org
>>>>> https://lists.lsst.org/mailman/listinfo/dm-users
>>>>>
>>>>> _______________________________________________
>>>>> dm-users mailing list
>>>>> dm-users at lists.lsst.org
>>>>> https://lists.lsst.org/mailman/listinfo/dm-users
>>>>
>>>
>>> _______________________________________________
>>> dm-users mailing list
>>> dm-users at lists.lsst.org
>>> https://lists.lsst.org/mailman/listinfo/dm-users
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.lsst.org/pipermail/dm-users/attachments/20150511/7dd8cf06/attachment-0001.html>
More information about the dm-users
mailing list