|
Hermes DRS
6.0
reduction software
|
|
Hermes DRS
6.0
reduction software
|
Computes the bias and the readout noise on the prescan region. More...
Public Member Functions | |
| def | __init__ |
| Make a Hermes component and initialize variables. | |
| def | optionsAreValid |
| checks user options | |
| def | __call__ |
| runs from here | |
 Public Member Functions inherited from hermes.pipeline.core.pipelinecomponent.PipelineComponent | |
| def | __init__ |
| def | __call__ |
| def | optionsAreValid |
| def | getOptions |
| def | setOptions |
| def | checkExistShapeUnits |
| def | isOption |
Public Attributes | |
| isDebug | |
| user may want intermediary result images in debug folder More... | |
| prescanRejectFactor | |
| reject factor | |
| prescanFilDim | |
| filter dimension | |
| constantBias | |
| constantBias can be "yes" or "no" or "model". More... | |
| results | |
| Public Attributes inherited from hermes.pipeline.core.pipelinecomponent.PipelineComponent | |
| options | |
| logger | |
| ConsoleLogSeverity | |
| default value | |
| canShowGraphics | |
| when machine does not run matplotlib. More... | |
Computes the bias and the readout noise on the prescan region.
PURPOSE :
Estimates bias level and read-out noise from analysis of prescan region.
Comparison with default values for the bias level and the read-out noise
are made. Note that the bias value in the prescan region is somewhat
different from that on the illuminated part of the detector (see procedures
to characterize the detector for the determination of the offset).
COMMENTS :
Outlying intensity values are counted and masked (in fact, they should
not occur in the prescan region; if there are too many outliers,
a warning will be generated).
The read-out noise is computed taking into account a possible (slight)
variation of the bias over the prescan region, by computation of a local
estimate of the bias.
IF the bias value is constant over the frame (as it should be),
the procedure would be more robust when using a global estimate of the
bias and not a local one. IF the bias value varies over the frame,
and this cannot be avoided by hardware intervention, the procedure would
be more robust when building a specific few-parameter model for the
shape of the variation of the bias over the frame, rather than
depending on a general local estimator for the bias value.
SYNTAX :
INPUT CONSTANTS FROM INSTRUMENT MODEL :
position of prescan region
defaultPrescanreadoutNoiseADU
INPUT DATA FROM RESULTDICT :
An observed image from self.results.rawImage (possibly the mean of a stack
of images, and in that case the number of images in stack is given
by "len(self.results["FITSFileNames"])").
The size of the input image is checked against the expected shape
of a raw image.
OPTIONS READ FROM OPTIONDICT :
MANDATORY :
OPTIONAL :
constantBias == "no" or "yes" or "model"
If absent, default value "yes" is used.
prescanFilDim
default value is 5
prescanRejectFactor
default value is 4.0
OUTPUT :
Values measured in the prescan region for bias and read-out noise, in
self.results.prescanBiasADU and self.results.prescanReadoutNoiseADU, also
written to self.results.header
TODO : comment on messages
in case of invalid input data or suspect frames | hermes.pipeline.components.computebiasprescan.computeBiasPrescan.constantBias |
constantBias can be "yes" or "no" or "model".
Default value is "yes".
| hermes.pipeline.components.computebiasprescan.computeBiasPrescan.isDebug |
user may want intermediary result images in debug folder
Determine the prescan area from the complete frame.
first get the dimensions Then compare with the passed image size if allowedSize.shape == self.results["rawImage"].shape and self.results["imageUnits"] == "ADU": determine local estimate of bias (median over size of filter) medfilt2d causes a memory crash. let's use gc to clean up memory before. Force writing of debug file as this solves the problem When not in debug mode, remove the unnecessary debug file we had to write
1.8.6