HAWKI Pipeline Reference Manual 1.8.12
Functions
Variance estimation for raw images.

Functions

cpl_image * hawki_image_create_variance (const cpl_image *image, double gain, double ron, int ndit, int ndsamples)
 Create the associated variance of a raw image.
cpl_imagelist * hawki_imglist_create_variances_and_delete (cpl_imagelist *imagelist_raw, double gain, double ron, int ndit, int ndsamples)
 Create the associated variance of raw imagelist.

Function Documentation

cpl_image* hawki_image_create_variance ( const cpl_image *  image,
double  gain,
double  ron,
int  ndit,
int  ndsamples 
)

Create the associated variance of a raw image.

Parameters:
imagethe raw image
gainthe gain of the detector (e-/ADU)
ronthe ron for one read of the detector (ADU)
nditthe number of total integrations.
ndsamplesthe number of non-destructive readouts.
Returns:
The variance image if suceed. NULL elsewhere.

The variance is computed with the following equation:

\[ \sigma^{2}(i,j) = \frac{6 S(i,j)}{5 g n_{nds} n_{dit}} (\frac{n_{nds}^{2} + 1}{n_{nds} + 1}) + \frac{12 RON^{2}(i,j)}{g^{2} n_{nds} n_{dit}} (\frac{n_{nds}-1}{n_{nds}+1})\]

where: $\sigma^{2}(i,j)$ is the total variance in (ADU $^{2}$), $S(i,j)$ is the signal in ADU of electrons in the pixel potencial well, so to the measured ADU we must first remove the bias, $g$ is the gain of the detector (in electrons/ADU), $n_{nds}$ is the number of non-denstructive reads (ndsamples), $n_{ndit}$ is the number of total integrations (ndit) and $RON(i,j)$ is the readout noise of a single readout of the detector (in ADU). Note that the dimensional inconsistency of the previous expression is only apparent. The expression used, however is only approximate, and is valid for RON << S(i,j). For low signal levels there may be significant deviations

Definition at line 81 of file hawki_variance.c.

Referenced by hawki_imglist_create_variances_and_delete().

cpl_imagelist* hawki_imglist_create_variances_and_delete ( cpl_imagelist *  imagelist_raw,
double  gain,
double  ron,
int  ndit,
int  ndsamples 
)

Create the associated variance of raw imagelist.

Parameters:
imagelist_rawthe raw imagelist
gainthe gain of the detector (e-/ADU)
ronthe ron for one read of the detector (ADU)
nditthe number of total integrations.
ndsamplesthe number of non-destructive readouts.
Returns:
The variance imagelist if suceed. NULL elsewhere.

The original imagelist_raw is empty after the process. This is so in order not to duplicate the amount of memory needed. However it will still have to be deleted with cpl_imagelist_delete. Also the returned pointer have to be deleted with cpl_imagelist_delete, of course.

The variance is computed with the following equation:

\[ \sigma^{2}(i,j) = \frac{6 S(i,j)}{5 g n_{nds} n_{dit}} (\frac{n_{nds}^{2} + 1}{n_{nds} + 1}) + \frac{12 RON^{2}(i,j)}{g^{2} n_{nds} n_{dit}} (\frac{n_{nds}-1}{n_{nds}+1})\]

where: $\sigma^{2}(i,j)$ is the total variance in (ADU $^{2}$), $S(i,j)$ is the signal in ADU of electrons in the pixel potencial well, so to the measured ADU we must first remove the bias, $g$ is the gain of the detector (in electrons/ADU), $n_{nds}$ is the number of non-denstructive reads (ndsamples), $n_{ndit}$ is the number of total integrations (ndit) and $RON(i,j)$ is the readout noise of a single readout of the detector (in ADU). Note that the dimensional inconsistency of the previous expression is only apparent. The expression used, however is only approximate, and is valid for RON << S(i,j). For low signal levels there may be significant deviations.

Definition at line 159 of file hawki_variance.c.

References hawki_image_create_variance().