Level1_ISS
Level 1 ISS ¶
- Level 1 ISS ¶
- !!!!!!!!!!!!!!!!ENTER CASSINI ISS JAVASCRIPT WIDGET!!!!!!!!!!!!!!!! ¶
- !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!ENTER JAVASCRIPT WIDGET!!!!!!!!!!!!!!!!!!!!!!!!!! ¶
ISS Radiometric Calibration ¶
The application cisscal applies the Cassini ISS radiometric calibration correction. ISIS3 cisscal was derived from the IDL 'CISSCAL' application developed by the Cassini Imaging Central Laboratory for Operations (CICLOPS). cisscal works for both the ISS Cameras; Narrow Angle ( NAC ) and Wide Angle ( WAC ).
It is recommeded to specifty the output calibrated image cube units to be in I/F Units and 32-bit for precision.
Example:
cisscal from=xxxx.cub to=xxxx_cal.cub fluxunits=I/F
Noise Removal ¶
Occasionally an amount of random noise remains in the ISS image after the calibration correction.
Remove random pixel noise ¶
Example:
noisefilter from=xxxx_cal.cub to=xxxx_stdz.cub toldef=stddev tolmin=2.5 tolmax=2.5 replace=null samples=5 lines=5
This example runs the noisefilter application, setting the boxcar size to 5x5 pixels and uses the standard deviation option, replacing the noise pixels with NULL.
- Thorough and accurate removal of random noise while retaining valid data can be a challenge. Initially, the process can be iterative in applying and evaluating results of noisefilter with various settings for boxcar size and tolerances. Often, noisefilter will be applied more than once with optimum settings to consecutive results of previous noisefilter runs.
- Replacing the noise with NULL at this stage is recommended in order to visually evaluate what is removed. The NULL-ed pixels can be filled in later with the average of surrounding valid pixels by using the low pass filter application ( lowpass ). See Filling in random NULL pixels .
- Replacing NULLs with the low pass filter application should be done BEFORE removing any frame-edge noise to avoid 'filling in' the 'trimmed' edges
Hit refresh if an image is not displayed in black box
Close up view of random pixel noise of an ISS-NAC image (Enceladus)
!!!!!!!!!!!!!!!!ENTER CASSINI ISS JAVASCRIPT WIDGET!!!!!!!!!!!!!!!! ¶
Fill-in NULL pixels ¶
- Overview of Spacial Filters: The Power of Spatial Filters
Random NULL pixels ¶
The resulting NULL-ed pixels from removing random pixel noise can be filled in with the average of surrounding valid pixels. This is achieved by applying the low pass filter application ( lowpass ) with specific parameters settings to replace only NULL pixels. See Fill-in Null Pixels .
- Filling in the random NULL pixels at this stage will also fill-in the truncated lines that contain NULLs as well.
- Replacing NULLs with the low pass filter application should be done BEFORE removing any frame-edge noise to avoid 'filling in' the 'trimmed' edges
Example for filling-in NULLS:
lowpass from=xxx_stdz.cub to=xxxx_fill.cub samples=3 lines=3 filter=outside null=yes hrs=no his=no lrs=no replacement=center
- If NULL pixels remain, lowpass can be applied consecutively on each output until the NULL's are completely filled with surrounding averages.
Truncated lines ¶
A number of ISS images have truncated lines. These truncated lines will contain valid data half way across in the sample direction with lines above and below containing valid data across the entire sample direction. The method to Fill-in Null Pixels using the ( lowpass ) application works nicely in this case.
- Filling in the random NULL pixels at this stage will also fill-in any random NULL pixels as well.
- Replacing NULLs with the low pass filter application should be done BEFORE removing any frame-edge noise to avoid 'filling in' the 'trimmed' edges.
Example:
 |
 |
|---|
Remove frame-edge noise ¶
Many of the ISS calibrated images will still contain invalid data along the edges within the first 1 to 2 pixels in the line and sample direction. This edge artifact might not be visibly noticable until the images are mosaicked together. This messy edge of the individual images will become evident as this invalid data overlaps valid data of a neighboring image.
Example:
trim from=xxxx_fill.cub to=_xxxx_trim.cub top=2 bottom=2 left=2 right=2
The trim application will replace the specified number of pixels along the edges with NULL values ( What Are Special Pixels? ).
Hit refresh if an image is not displayed in black box
Close up view of upper left hand corner of an ISS-NAC image (Enceladus)
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!ENTER JAVASCRIPT WIDGET!!!!!!!!!!!!!!!!!!!!!!!!!! ¶
Before/After trimming off edge noise
- Be aware that any boxcar filtering that includes modification of NULL values, the resulting trimmed (NULL-ed) edges will also be modified.
DropLines_sm.png View (162 KB) Makayla Shepherd, 2016-06-01 09:05 AM
DropLinesFilled_sm.png View (149 KB) Makayla Shepherd, 2016-06-01 09:05 AM