Input is a single file in either BSQ or BIL format, containing at least 3 channels. The user may specify which channels are to be used, by means of the USEBANDS parameter.
OUT contains the name of the output dataset that contains the transformed image. If no output dataset is specified, the statistics will be gathered and reported, but no transformation will be applied.
The standard Vicar size field ( starting_line, starting_sample, number_of_lines, number_of_samples).
Starting line of the portion of the image that you wish to process.
Starting sample of the portion of the image that you wish to process.
Number of lines in the portion of the image that you wish to process.
Number of samples in the portion of the image that you wish to process.
If the value of the parameter MATRIX is "COV" (the default), the image's covariance matrix is used to determine the principal components. If the value of MATRIX is "CORR", then the correlation matrix is used instead.
Statistics are gathered using only every n'th line and n'th sample of the image, or region of interest (AREA) within the image. The parameter INC specifies the value on "n".
If this parameter is defaulted, all of the channels in the input file are used in the transformation. If the user wishes to use only a subset of the channels, the USEBANDS parameter is used to select the channels to be used.
If this parameter is defaulted, the number of output components (channels) will be equal to the number of input channels used. If the user wishes to output only a subset of the principal components, the OUTPUT_PCS parameter is used to select the principal components to be output.
Sets of (Starting_line, Starting_sample, Number_of_lines, Number_of_samples) are given to define subareas used to generate the image statistics. Up to 50 set of subareas may be supplied. The default is that the entire image is sampled.
A rescaling factor is included in the overall transformation to reposition the output values in a range appropriate for output data type (BYTE, HALF, FULL, REAL). The MEAN parameter specifies the desired mean value for the output image channels. If MEAN is defaulted, a value of 127.5 is used. If the AREA parameter has been used, this target mean is for the AREA(s) of interest only. To suppress rescaling entirely, assign the parameter SIGMA a negative value.
A rescaling factor is included in the overall transformation to reposition the output values in a range appropriate for the data output. The SIGMA parameter specifies the desired standard deviation from the mean value in each of the output image channels. If SIGMA is defaulted, then a value of 45.0 is used. If SIGMA is assigned a negative value, the rescaling operation is suppressed. If the AREA parameter has been used, this target standard deviation is for the AREA(s) of interest only.
A pixel is excluded from use in determining the transformation if the DN is equal to 'EXCLUDE' in each channel.
The value of the PRINT variable determines the detail of the printed
information output to the standard output and session log.
If ALL (the default), the complete list of input means, standard deviations,
interchannel covariances and correlations (if computed), eigenvalues,
eigenvectors, and transformation equations is printed.
If SUMMARY, these tables of values are only summarized, and the eigenvectors
and transformation equations are omitted.
If NOPRINT, only the EIGEN version number is printed.
If the XFORM_PARM parameter is given a value, the transformation matrix will be saved as a VICAR parameter dataset, and have the dataset name specified by the save parameter. This parameter dataset may then be included in the parameters for XFORM, to repeat this transformation on other datasets.