STScI Logo

xyztoim toolbox.imgtools



xyztoim -- Interpolate table values, writing results to an image.


xyztoim intable output


This routine reads tables -- or text files in row/column format -- and writes 1-D or 2-D output images. Each input table or file contains columns of X, Y (if 2-D), and Z values. This task fits a surface to Z as a polynomial function of X and Y and evaluates the fit at each pixel of the corresponding output image. The function that is to be fit may be expressed as a Chebyshev or Legendre polynomial or as an ordinary power series. The IRAF gsurfit package is used to fit and evaluate the function.

A linear relationship between X & Y and pixel numbers is assumed. X and Y cannot be right ascension and declination, for example. unless the declination is small.

Header parameters are added to the output images to give some information on the fit.


intable [file name template]
List of input tables or text files. See also xname, yname, zname, which are used to specify the column names. The same names are used for all input tables.
output [file name template]
List of output 1-D or 2-D images. The number of output images must be the same as the number of input tables or files. All images will be 1-D or all will be 2-D, depending on whether yorder is zero or not.
(xname = "c1") [string]
Name of column for X values. This parameter is used to specify which column in the input table contains the X values. X and Y are the independent variables (or just X for 1-D), and Z is the dependent variable. The default of "c1" is appropriate for the case that intable is a text file with X values in the first column. If intable lists more than one file name, the same column names are used for all files.
(yname = "c2") [string]
Name of column for Y values. For a one-dimensional fit and 1-D output image, you can either set yorder to zero or set yname to null or blank. Note that zname is the column name for the dependent variable regardless of whether a 1-D or 2-D fit is desired. See also the description for xname.
(zname = "c3") [string]
Name of column for Z values. This column contains the dependent variable values. For a 1-D fit, values should be specified for xname and zname, and yname will not be used. See also the description for xname.
(nx = 512) [integer, min=1, max=INDEF]
Width of image, in pixels.
(ny = 512) [integer, min=1, max=INDEF]
Height of image, in pixels. This is ignored if yorder is zero, i.e. for a 1-D fit.
(xorder = [integer, min=1, max=INDEF]
Number of coefficients for function in X. This number does not include coefficients for cross terms. For example, if xorder and yorder are both equal to two, the function will be
c+ c2*X + c3*Y if cross_terms = no,

and it will be

c+ c2*X + c3*Y + c4*X*Y if cross_terms = yes,

where c1, c2, c3, and c4 are the coefficients of the fit.

(yorder = [integer, min=0, max=INDEF]
Number of coefficients for function in Y. For a one-dimensional fit and 1-D output image, set yorder to zero.
(x= INDEF) [real]
X value at first pixel. If x1 is INDEF, the minimum X value in the input table will be used. The parameters x1, x2, y1 and y2 serve two purposes. They specify the range over which the fit is to be performed, and they specify the values of the independent variables at the corners of the image.
(x= INDEF) [real]
X value at last pixel. If x2 is INDEF, the maximum X value in the input table will be used.
(y= INDEF) [real]
Y value at first pixel. If y1 is INDEF, the minimum Y value in the input table will be used. In the 1-D case (i.e. if yorder is zero), y1 and y2 are ignored.
(y= INDEF) [real]
Y value at last pixel. If y2 is INDEF, the maximum Y value in the input table will be used.
(cross_terms = yes) [boolean]
Include cross-terms? If this is set to no, the function will consist of the sum of a polynomial in X and a polynomial in Y. If cross_terms = yes, the function can include terms such as X*Y or (X**2)*Y.
(function = "chebyshev") [string]
[allowed values: chebyshev | legendre | polynomial]

Function to be fit. The default value of "chebyshev" is almost always appropriate. Numerical roundoff may be severe if "polynomial" is selected, so this choice is not recommended.

(verbose = yes) [boolean]
Print the names of the input table and output image?
(coefficients = no) [boolean]
Print the coefficients? The coefficients are not printed in user-friendly format. They are the values returned by the dgssave subroutine in gsurfit. The first eight numbers describe the fit (e.g. xorder, yorder), and the remaining values are the coefficients. These may be passed to the dgsrestore subroutine in order to restore these coefficients to a gsurfit structure.


1. Suppose the input file "test.lis" contains X, Y, and Z values in the first three columns as follows:

 #  x    y    z
    0.   0.   0.
    1.   0.   1.
    0.   1.   2.

Create a 4 by 4 image "test" with X and Y values starting at zero at pixel (1,1) and increasing by 0.5 per pixel. Fit a plane, i.e. polynomial with two coefficients for each axis and no cross terms. We can use the default values for column names and order of fit.

    im> xyztoim test.lis test nx=4 ny=4 x1=0 y1=0 x2=1.5 y2=1.5 \
    >>> cross_terms=no function="polynomial"
    test.lis --> test;  rms = 2.31111E-16

The RMS error in the fit is machine roundoff in this case. Use listarea to examine the pixel values:

    fo> listarea test[*,-*]
    Image:  test[*,-*]
     Sample    1     2     3     4
          4  3.0   3.5   4.0   4.5 
          3  2.0   2.5   3.0   3.5 
          2  1.0   1.5   2.0   2.5 
          1  0.0   0.5   1.0   1.5 

2. File "1d.lis" contains the following:

	#  z = 0.1*x**2 - 2*x + 5
	# x   z
 	-5  17.5
 	-4  14.6
 	-3  11.9
 	-2   9.4
 	-1   7.1
  	 0   5.0
  	 1   3.1
  	 2   1.4
  	 3  -0.1
  	 4  -1.4
  	 5  -2.5
  	 6  -3.4

Create a 1-D output image as follows.

    im> xyztoim 1d.lis 1d.hhh xname="c1" zname="c2" nx=12 \
    >>> xorder=3 yorder=0
    1d.lis --> 1d.hhh;  rms = 1.84047E-15

The data values of "1d.hhh" will be the same as "1d.lis", and the header will contain:

    CRPIX1  =                   1.
    CRVAL1  =                  -5.
    CTYPE1  = 'PIXEL   '
    CD1_1   =                   1.
    XORDER  =                    3
    RMSERR  =  1.8404672640492E-15
    HISTORY   input table name 1d.lis
    HISTORY   column names c1 c2
    HISTORY   a Chebyshev function was fit



This task was written by Phil Hodge.


xyztable, gsurfit

Source Code · Package Help · Search Form · STSDAS