SMSPEC

Smooth spectra


EnvironmentsPYTHON :: EASI :: MODELER
Batch ModeYes
Quick linksDescription :: Parameters :: Parameter descriptions :: Details :: Acknowledgements :: References :: Related

Back to top

Description


SMSPEC smooths the band spectra in an image data set or smooths the spectra from a spectra library file. An image data set must contain center-wavelength metadata.
Back to top

Parameters


Name Type Length Value range
Input File Name String 0 -    
Input Spectral Library File String 0 -    
Output Spectral Library File String 0 -    
Spectral Library File Format String 0 - 1 SPCB | XLS | TXT
Default: SPCB
Valid Bands Only String 0 - 1 YES | NO
Default: NO
Spectrum Filter Type String 0 - 1 GAU | MED
Default: GAU
Filter Width Float 0 - 1 0.0 -
Default: 30.0
Wavelength List 1 Float 0 - 16  
Wavelength List 2 Float 0 - 16  
Back to top

Parameter descriptions

Input File Name

Specifies the name of the file that contains the data set to smooth.

The stored pixel values are smoothed; radiometric transformations in the image metadata are ignored.

If an input (SPECFILI) or output (SPECFILO) spectral library file is specified, this parameter cannot be specified.

Input Spectral Library File

Specifies the name of the spectral library file that contains the spectra to be smoothed.

If an input file (FILE) is specified, this parameter cannot be specified.

Output Spectral Library File

Specifies the name of the spectral library file to be created to receive the smoothed sepctra.

If an input file (FILE) is specified, this parameter cannot be specified.

The following rules apply to the output spectral library file extensions:

Spectral Library File Format

Specifies the file format of the output spectral library file.

Supported formats are:
Note: If the "TXT" type is specified, the function also generates a file with a ".gav" extension; this file contains the schema for the file.

Valid Bands Only

Specifies whether the selected bands are to be restricted to those with 'plot' or 'bmask' (begin mask) quality values.

Available options are:

This parameter has no effect if the input file contains no band-validity metadata.

Note: Either upper- or lower-case characters may be used in the assigned strings.

Spectrum Filter Type

Specifies the type of filter to use for filtering the band vectors.

Supported values are:

Filter Width

Specifies the filter width, in nanometers. The default is 30nm. For a Gaussian convolution filter, the width is the standard deviation of the Gaussian function.

Wavelength List 1

Specifies, in nanometers, the start of each wavelength interval within which smoothing is to be performed.

If this parameter is not specified, the spectra are smoothed across their full wavelength range.

If this parameter is specified, WLEN2 must also be specified.

Wavelength List 2

Specifies, in nanometers, the end of each wavelength interval within which smoothing is to be performed.

If this parameters is not specified, the spectra are smoothed across their full wavelength range

If this parameter is specified, WLEN1 must also be specified.

Back to top

Details

Spectrum values within the wavelength intervals specified with WLEN1 and WLEN2 (or across the full wavelength range of the spectra) are smoothed either by convolution with a one-dimensional Gaussian kernel (SPFILT="GAU") or by applying a one-dimensional median filter (SPFILT="MED"). The width of the filter is defined by the FWIDTH parameter; in general, the larger the filter, the greater the degree of smoothing.

In the case of a Gaussian convolution filter, the value of FWIDTH is the standard deviation of the Gaussian function. In the case of a median filter, FWIDTH is the overall neighborhood dimension.

SMSPEC may be used to subdue artifacts in reflectance images or spectra caused by some form of radiometric distortion in the corresponding radiance images or spectra before atmospheric correction. An example of such distortion is spectral line curvature, or 'smile'.

If the noise to be removed from the spectra is in the form of sharp impulses or spikes, the median filter may suppress them while causing less distortion than the Gaussian filter. If the noise is in the form of relatively broad ripples, the Gaussian filter may produce a more acceptable result.

If SMSPEC is applied to an image data set (to smooth along the band dimension), it smoothes the stored pixel values, not the values resulting from the application of any radiometric transformation stored in the image metadata.

Back to top

Acknowledgements

PCI Geomatics received financial support from the Canadian Space Agency/L'Agence Spatiale Canadienne through the Earth Observation Application Development Program (EOADP) for the development of this software, under contract 9F028-0-4914/09.

Back to top

References

Landgrebe, D.A., "Information Extraction Principles and Methods for Multispectral and Hyperspectral Image Data," Chapter 1 of Information Processing of Remote Sensing, World Scientific Publishing Co., River Edge N.J.

Szeredi T., Staenz K., Neville R.A., Canada Centre for Remote Sensing, 1999, "Automatic Endmember Selection: Part I, Theory," submitted to Remote Sensing of the Environment, Elsevier Science, Feb. 1999

© PCI Geomatics Enterprises, Inc.®, 2024. All rights reserved.