JavaScript is not enabled, using print styling. Please note, this may disable the navigation menu.

specdivfilter

Apply a spectral division filter

Overview
Parameters
Example 1
Example 2

This program applies a spectral division filter to a cube. A spectral filter works between bands as opposed to a spacial operation on a single band. As a division filter, it divides the original pixel by the average.

Related Applications to Previous Versions of ISIS

This program replaces the following application existing in previous versions of ISIS:

boxfilter

History

Stacy Alley	2008-09-02	Original version
Mackenzie Boyd	2009-06-05	Modified documentation, added error checking for the number of bands, and added examples

Parameters

Files

Input cube to filter

Type	cube
File Mode	input
Filter	*.cub

The resultant filtered cube

Type	cube
File Mode	output
Pixel Type	real
Filter	*.cub

Boxcar Size

This is the total number of bands in the boxcar. It must be odd and can not exceed twice the number of bands in the cube. In general, the size of the boxcar does not cause the program to operate significantly slower.

Type	integer
Minimum	1 (inclusive)
Odd	This value must be an odd number

Boxcar Settings

Valid minimum pixel value that will be used in boxcar computation. If a pixel value is less than LOW then it will not be used when computing the average.

Type	double
Internal Default	Use all pixels
Less Than	HIGH

Valid maximum pixel value that will be used in boxcar computation. If a pixel value is greater than HIGH then it will not be used when computing the average.

Type	double
Internal Default	Use all pixels
Greater Than	LOW

Example 1

Example of usage of the spectral division filter.

This example shows a spectral division filter being applied to the image with a 15 depth boxcar. This cube has 256 bands so 15 is relatively small. No averaging is done between bands, the divisoin filter more increases differences between bands. Here is an image which highlights which pixels a single pass would influence. The image shows a 9x9 dot instead of a single pixel highlighted to make it more visible. In this picture the boxcar would be seven or more bands, with the output pixel being one of the visible pixels. Depending on how large the boxcar was many of the values taken in could not be original, at a minimum, one value will be mirrored since there are only six available bands.

Command Line

This example will filter the image using 15 bands and no limits on high or low.

specdivfilter 
          from=/work1/mboyd/CM_1514302573_1.cub to=/work1/mboyd/postlow.cub 
          bands=15

GUI Screenshot

Example GUI

This program's GUI

Screenshot of the GUI with parameters set to perform the specdivision filter with a 15 band boxcar on the image CM_1514302573_1.ir.cub.

Input Image

Input image spectral graph before specdivfilter

The image graph before the filter

Parameter Name: FROM

This is the spectral plot of the center point of the image, (32, 32), and shows all bands, 1- 256. No filter has been applied, and the graph clearly shows how the value of line 32 sample 32 differs from band to band.

Output Image

Output image spectral graph after being filtered

The image graph after the filter

Parameter Name: TO

This is the image after the filter. First note the change in scale, everything has been amplified. Even in the apparent flat area small differences have been amplified.

Example 2

Example of usage of the spectral division filter.

This example shows a spectral division filter being applied to the image with a 61 depth boxcar. This cube has 256 bands so 61 takes in over a fifth. In this case an image begins to show through, however, this is not because of average but because of amplificatino of differences from the original.