vircam_stats
[Groups of Support Routines]

---

Detailed Description

These are support routines used for doing statistics on data arrays.

Author:

Jim Lewis, CASU, CASU

Functions
float	vircam_med (float *data, unsigned char *bpm, long npts)
double	vircam_dmed (double *data, unsigned char *bpm, long npts)
float	vircam_mean (float *data, unsigned char *bpm, long npts)
double	vircam_dmean (double *data, unsigned char *bpm, long npts)
int	vircam_meansig (float *data, unsigned char *bpm, long npts, float *mean, float *sig)
int	vircam_meansigcut (float *data, unsigned char *bpm, long npts, float lcut, float hcut, float *mean, float *sig)
void	vircam_qmedsig (float *data, unsigned char *bpm, long npts, float thresh, int niter, float lowv, float highv, float *median, float *sigma)
void	vircam_medmad (float *data, unsigned char *bpm, long np, float *med, float *mad)
void	vircam_medmadcut (float *data, unsigned char *bpm, long np, float lcut, float hcut, float *med, float *mad)
void	vircam_medsig (float *data, unsigned char *bpm, long np, float *med, float *sig)
int	vircam_sumbpm (unsigned char *bpm, int npts, int *sumb)

---

Function Documentation

double vircam_dmean	(	double *	data,
		unsigned char *	bpm,
		long	npts
	)

Name:

vircam_dmean

Purpose:

Find the mean of a double array with a possible bad pixel mask

Description:

If a bad pixel mask is not included a straight forward mean is performed on the input data array. If a bad pixel mask is included then the data good data are copied into a temporary array and then the mean is performed. This is a routine for double precision data.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	npts	Number of pixels in the data

Return values:

	mean	When all goes well
	CX_MAXDOUBLE	if all are flagged as bad

Author:

Jim Lewis, CASU

Definition at line 296 of file vircam_stats.c.

double vircam_dmed	(	double *	data,
		unsigned char *	bpm,
		long	npts
	)

Name:

vircam_dmed

Purpose:

Find the median of a double array with a possible bad pixel mask

Description:

If a bad pixel mask is not included a straight forward median is performed on the input data array. If a bad pixel mask is included then the data good data are copied into a temporary array and then the median is performed. This is a routine for double precision data.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	npts	Number of pixels in the data

Return values:

	median	When all goes well
	CX_MAXDOUBLE	if all are flagged as bad

Author:

Jim Lewis, CASU

Definition at line 165 of file vircam_stats.c.

Referenced by vircam_platesol().

float vircam_mean	(	float *	data,
		unsigned char *	bpm,
		long	npts
	)

Name:

vircam_mean

Purpose:

Find the mean of a float array with a possible bad pixel mask

Description:

If a bad pixel mask is not included a straight forward mean is performed on the input data array. If a bad pixel mask is included then the data good data are copied into a temporary array and then the mean is performed. This is a routine for floating point data.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	npts	Number of pixels in the data

Return values:

	mean	When all goes well
	CX_MAXFLOAT	if all are flagged as bad

Author:

Jim Lewis, CASU

Definition at line 241 of file vircam_stats.c.

Referenced by vircam_mkconf().

int vircam_meansig	(	float *	data,
		unsigned char *	bpm,
		long	npts,
		float *	mean,
		float *	sig
	)

Name:

vircam_meansig

Purpose:

Find the mean and sigma of a floating point array with a possible bad pixel mask

Description:

A straight-forward mean and standard deviation calculation is done. If a bad pixel mask is included, then the flagged pixels are removed from the calculations.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	npts	Number of pixels in the data
	mean	The output mean
	sig	The output dispersion

Return values:

	VIR_OK	If all went well
	VIR_WARN	If the variance is negative. Sigma is passed back as CX_MAXFLOAT. Also if all data points are flagged. In that case both the mean and sigma are set to CX_MAXFLOAT.

Author:

Jim Lewis, CASU

Definition at line 357 of file vircam_stats.c.

Referenced by vircam_difference_image().

int vircam_meansigcut	(	float *	data,
		unsigned char *	bpm,
		long	npts,
		float	lcut,
		float	hcut,
		float *	mean,
		float *	sig
	)

Name:

vircam_meansigcut

Purpose:

Find the mean and sigma of a floating point array with a possible bad pixel mask and with lower and upper cuts.

Description:

A straight-forward mean and standard deviation calculation is done. If a bad pixel mask is included, then the flagged pixels are removed from the calculations. Data lower than the lower cut and data higher than the upper cut are removed also.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	npts	Number of pixels in the data
	lcut	Lower value for data window
	hcut	Upper value for data window
	mean	The output mean
	sig	The output dispersion

Return values:

	VIR_OK	If all went well
	VIR_WARN	If the variance is negative. Sigma is passed back as CX_MAXFLOAT. Also if all data points are flagged. In that case both the mean and sigma are set to CX_MAXFLOAT.

Author:

Jim Lewis, CASU

Definition at line 446 of file vircam_stats.c.

Referenced by vircam_illum(), and vircam_photcal().

float vircam_med	(	float *	data,
		unsigned char *	bpm,
		long	npts
	)

Name:

vircam_med

Purpose:

Find the median of a float array with a possible bad pixel mask

Description:

If a bad pixel mask is not included a straight forward median is performed on the input data array. If a bad pixel mask is included then the data good data are copied into a temporary array and then the median is performed. This is a routine for floating point data.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	npts	Number of pixels in the data

Return values:

	median	When all goes well
	CX_MAXFLOAT	if all are flagged as bad

Author:

Jim Lewis, CASU

Definition at line 89 of file vircam_stats.c.

Referenced by vircam_backmap(), vircam_destripe(), vircam_jmp_interleave(), vircam_matchxy(), vircam_medmad(), vircam_medmadcut(), and vircam_medsig().

void vircam_medmad	(	float *	data,
		unsigned char *	bpm,
		long	np,
		float *	med,
		float *	mad
	)

Name:

vircam_medmad

Purpose:

Find the median and median absolute deviation in a data array

Description:

The median of an array is found, taking into account any flagged values. An array of absolute deviations is created and the median is found of that array.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	np	Number of pixels in the data
	med	The output median
	mad	The output median absolute deviation

Returns:

Nothing

Author:

Jim Lewis, CASU

Definition at line 620 of file vircam_stats.c.

References vircam_med().

Referenced by vircam_defringe(), vircam_difference_image(), vircam_genbpm(), vircam_illum(), vircam_matchstds(), and vircam_photcal().

void vircam_medmadcut	(	float *	data,
		unsigned char *	bpm,
		long	np,
		float	lcut,
		float	hcut,
		float *	med,
		float *	mad
	)

Name:

vircam_medmadcut

Purpose:

Find the median and median absolute deviation in a data array with a high and/or low cut.

Description:

The median of an array is found, taking into account any flagged values and any values outside a low and high data value cut. An array of absolute deviations is created and the median is found of that array.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	np	Number of pixels in the data
	lcut	Lower value for data window
	hcut	Upper value for data window
	med	The output median
	mad	The output median absolute deviation

Returns:

Nothing

Author:

Jim Lewis, CASU

Definition at line 679 of file vircam_stats.c.

References vircam_med().

void vircam_medsig	(	float *	data,
		unsigned char *	bpm,
		long	np,
		float *	med,
		float *	sig
	)

Name:

vircam_medsig

Purpose:

Find the median and sigma about that median.

Description:

The median of an array is found, taking into account any flagged values. The sigma is then worked out about the median.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask or NULL
	np	Number of pixels in the data
	med	The output median
	sig	The output sigma

Returns:

Nothing

Author:

Jim Lewis, CASU

Definition at line 746 of file vircam_stats.c.

References vircam_med().

Referenced by vircam_destripe(), vircam_genbpm(), and vircam_illum().

void vircam_qmedsig	(	float *	data,
		unsigned char *	bpm,
		long	npts,
		float	thresh,
		int	niter,
		float	lowv,
		float	highv,
		float *	median,
		float *	sigma
	)

Name:

vircam_qmedsig

Purpose:

Find a "quick" median and sigma of a floating point array with a bad pixel mask, with lower and upper cuts and with clipping.

Description:

Data outside the data window and data flagged in the bad pixel mask are removed from further consideration. The remaining data are histogrammed in 1 adu bins. A number of iterations are used to find the median and sigma by finding the meidan in the histogram and its quartiles, then using these in conjunction with user defined clipping thresholds to define new estimates.

Language:

C

Parameters:

	data	Input data
	bpm	Input bad pixel mask
	npts	Number of pixels in the data
	thresh	Clipping threshold
	niter	Number of clipping iterations
	lowv	Lower value for data window
	highv	Upper value for data window
	median	The output median
	sigma	The output dispersion

Returns:

Nothing

Author:

Jim Lewis, CASU

Definition at line 537 of file vircam_stats.c.

Referenced by vircam_jmp_skycor().

int vircam_sumbpm	(	unsigned char *	bpm,
		int	npts,
		int *	sumb
	)

Name:

vircam_sumbpm

Purpose:

Get the sum from a bad pixel mask

Description:

Sum all the values in a bad pixel mask. This is useful for telling how many pixels are bad.

Language:

C

Parameters:

	bpm	Input bad pixel mask
	npts	Number of pixels in the bpm
	sumb	The output sum

Return values:

VIR_OK

Always

Author:

Jim Lewis, CASU

Definition at line 808 of file vircam_stats.c.

---