Skip to content

Scripts for Cassini, Juno, Voyager, & Galileo Imagery Processing

Notifications You must be signed in to change notification settings

Digus12/cassini_processing

 
 

Repository files navigation

Cassini, Voyager, & Galileo Image Processing

Scripts and utilities for processing Juno, Cassini, Voyager, & Galileo imagery from NASA PDS archived data.

Overview

This is a collection of Python and shell scripts for processing Cassini and Voyager imagery from their PDS archived IMG and IMQ files into both ISIS3 cube files and TIFFs. The scripts require a working and initialized installation of the USGS ISIS3 software from https://isis.astrogeology.usgs.gov/. Ensure the ISIS3 mission kernels are all installed.

These alone do not produce the lovely full-color images you see as the finished products. Images output by these scripts would likely require additional work in Photoshop (or your favorite photo editing software). A working knowledge of ISIS3 also comes in handy.

Mission-Specific Documentation

Scripts

initcass.sh

Performs initalization of ISIS3 and adds these scripts to your shell's search path. EDIT this file before using so it reflects your installation environment.

To Use, 'script' it:

. /path/to/initcass.sh

process.py

Performs the nessessary steps for converting a PDS IMG file into a calibrated Tiff file. Tiff files are created as unsigned 16 bit grayscale. Files are output in a [PRODUCT_ID]_[TARGET]_[FILTER 1]_[FILTER 2]_[IMAGE DATE/TIME].(cub|tif) format (i.e. "N1674923569_SATURN_HAL_CL2_2011-01-28_15.45.16.tif").

usage: process.py [-h] -d DATA [DATA ...] [-m] [-f FILTER [FILTER ...]]
                  [-t TARGET] [-s] [-v] [-w WIDTH [WIDTH ...]]
                  [-H HEIGHT [HEIGHT ...]] [-S] [-p PROJECTION] [-n]
                  [-o OPTION [OPTION ...]]

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        Source PDS dataset
  -m, --metadata        Print metadata and exit
  -f FILTER [FILTER ...], --filter FILTER [FILTER ...]
                        Require filter or exit
  -t TARGET, --target TARGET
                        Require target or exit
  -s, --skipexisting    Skip processing if output already exists
  -v, --verbose         Verbose output (includes ISIS3 command output)
  -w WIDTH [WIDTH ...], --width WIDTH [WIDTH ...]
                        Require width or exit
  -H HEIGHT [HEIGHT ...], --height HEIGHT [HEIGHT ...]
                        Require height or exit
  -S, --skipspice       Skip spice initialization
  -p PROJECTION, --projection PROJECTION
                        Map projection (Juno)
  -n, --nocleanup       Don't clean up, leave temp files
  -o OPTION [OPTION ...], --option OPTION [OPTION ...]
                        Mission-specific option(s)

Mission-Specific Options:

  • Cassini:

    ringplane=true|false - Optionally treats an image using ring-plane geometry

  • Juno:

    vt=n - Apply vertical (top & bottom) trimming on each framelet where n is the number of pixels trimmed. histeq=true|false - Optionally run histogram equalization on the output images

Examples:

Convert file 'N1674923569_1.LBL' to tiff:

process.py -d N1674923569_1.LBL

Print information about 'N1674923569_1.LBL' and exit:

process.py -d N1674923569_1.LBL -m

Convert file 'N1674923569_1.LBL' to tiff, but only if it points at Saturn:

process.py -d N1674923569_1.LBL -t SATURN

Convert file 'N1674923569_1.LBL' to tiff, but only if it utilized either CB2 or MT2 filters:

process.py -d N1674923569_1.LBL -f CB2 MT2

Convert file 'N1674923569_1.LBL' to tiff, but only if it hasn't been run before:

process.py -d N1674923569_1.LBL -s

Convert file 'N1674923569_1.LBL' to tiff using ringplane georeferencing:

process.py -d N1674923569_1.LBL -r

Convert an entire directory, limiting to Enceladus images taken in RED, GRN, and BL1:

process.py -d *.LBL -t ENCELADUS -f RED GRN BL

Note: The Voyager process includes automatically addressing Reseaus using the ISIS3 findrx and remrx commands. These are imperfect, but still improve the image.

Without reseau removal:

With reseau removal:

match.py

Computes min/max values for a group of cube files and exports them to tiff file with a matching stretch. This is used to ensure a correct luminance across filters.

usage: match.py [-h] -d DATA [DATA ...] [-f FILTERS [FILTERS ...]]
                [-t TARGETS [TARGETS ...]]

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        Source PDS dataset(s)
  -f FILTERS [FILTERS ...], --filters FILTERS [FILTERS ...]
                        Require filter(s) or exit
  -t TARGETS [TARGETS ...], --targets TARGETS [TARGETS ...]
                        Require target(s) or exit

Output Example:

Once mosaicked, brightness matched images will blend without requiring post-processing adjustments.

Examples:

Process all the cubes in the working directory:

match.py -d *.cub

Process a set of specific files:

match.py -d N1684428714_TITAN_CL1_GRN_2011-05-18_16.03.21.cub N1684686374_TITAN_RED_CL2_2011-05-21_15.37.40.cub N1684686408_TITAN_CL1_GRN_2011-05-21_15.38.14.cub 

Process all the cubes in the working directory that target Titan:

match.py -d *.cub -t titan

Process all the cubes in the working directory that target Titan and use RED/GRN/BL1 filters:

match.py -d *.cub -t titan -f RED GRN BL1

get_coiss.sh

Simple script to fetch ISS archives. Specified by archive number.

For example, to download coiss_2099.tar.gz use:

get_coiss.sh 2099

compose_rgb.py

Process three PDS files and compose them into a single color image.

usage: compose_rgb.py [-h] -r RED -g GREEN -b BLUE [-m] [-v]

optional arguments:
  -h, --help            show this help message and exit
  -r RED, --red RED     Data label for the red channel
  -g GREEN, --green GREEN
                        Data label for the green channel
  -b BLUE, --blue BLUE  Data label for the blue channel
  -m, --match           Force matching stretch values
  -v, --verbose         Verbose output (includes ISIS3 command output)

Examples:

Process three files and create a RGB output using shared stretch min/max values:

compose_rgb.py -r N1713218719_1.LBL -g N1713218686_1.LBL -b N1713218653_1.LBL -m

Output:

Process three files and create a RGB output, each channel independently stretched:

compose_rgb.py -r N1713218719_1.LBL -g N1713218686_1.LBL -b N1713218653_1.LBL

Output:

cissident.py

Prints out image metadata for a list of label files

usage: cissident.py [-h] [-d DATA [DATA ...]]

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        PDS Label Files

Examples:

cissident.py -d *.LBL

Prints output similar to the following, with columns as Target, Filter #1, Filter #2, Image Time, Width, Height, Bits per pixel, Camera, and File.

                    TITAN|  CL1|  CB3|   2011-04-20_12.28.45| 1024| 1024|  16|  Narrow| N1681996622_1.LBL
                    TITAN|  CL1|  MT1|   2011-04-20_12.45.45| 1024| 1024|  16|  Narrow| N1681997642_1.LBL
                    TITAN|  CB3|  CL2|   2011-04-19_11.56.10| 1024| 1024|   8|    Wide| W1681908267_1.LBL
                    TITAN|  CL1|  VIO|   2011-04-19_11.56.58| 1024| 1024|   8|    Wide| W1681908315_1.LBL
                    TITAN|  CL1|  BL1|   2011-04-19_11.57.31| 1024| 1024|   8|    Wide| W1681908348_1.LBL
                    TITAN|  CL1|  GRN|   2011-04-19_11.58.04| 1024| 1024|   8|    Wide| W1681908381_1.LBL
                    TITAN|  CL1|  RED|   2011-04-19_11.58.37| 1024| 1024|   8|    Wide| W1681908414_1.LBL
                    TITAN|  CB2|  CL2|   2011-04-19_12.13.44| 1024| 1024|   8|    Wide| W1681909321_1.LBL

getmodel.py

Downloads a simulated view of the specified image from http://space.jpl.nasa.gov/

usage: getmodel.py [-h] -d DATA [-f FOV] [-p PCT]

optional arguments:
  -h, --help            show this help message and exit
  -d DATA, --data DATA  Source dataset to model
  -f FOV, --fov FOV     Field of view (angle)
  -p PCT, --pct PCT     Body width as percentage of image

Examples:

Fetch a simulated image for a specified dataset with defaults selected:

getmodel.py -d W1692470929_1.LBL

Original Image:

Modeled Image:

Fetch a simluated image using a custom field of view (degrees):

getmodel.py -d W1692470929_1.LBL -f 40

cub2tiff.py

Converts a Cassini ISS cube file into a 16 bit (unsigned integer) tiff.

usage: cub2tiff.py [-h] -d DATA [DATA ...]

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        Source cube file

project.py

Converts an input cube from the original camera geometry to a map projected format.

usage: project.py [-h] -d DATA [DATA ...] [-p PROJECTION] [-m MAP]

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        Source cube file
  -p PROJECTION, --projection PROJECTION
                        Desired map projection
  -m MAP, --map MAP     Input map

Examples:

Project an input cube to equirectangular, preserving the cube's resolution.

project.py -d 1173J2-002_Vg2_CALLISTO_CLEAR_1979-07-08_14.06.23.cub -p equirectangular

Project an input cube to match the projection and resolution of another cube.

project.py -d 1173J2-002_Vg2_CALLISTO_CLEAR_1979-07-08_14.06.23.cub -m map.cub

matchmap.py

Batch processes input cubes to match the projection and resolution of another map projected input cube.

usage: matchmap.py [-h] -d DATA [DATA ...] [-m MAP] -o OUTPUT

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        Source cube file
  -m MAP, --map MAP     Input map
  -o OUTPUT, --output OUTPUT
                        Output Directory

Examples:

Project a cube to equirectangular, then process all the cubes to match.

project.py -d 1173J2-002_Vg2_CALLISTO_CLEAR_1979-07-08_14.06.23.cub -p equirectangular
mv 1173J2-002_Vg2_CALLISTO_CLEAR_1979-07-08_14.06.23_equirectangular.cub map.cub
mkdir remapped
matchmap.py -d *CALLISTO*cub -m map.cub -o remapped

matchcam.py

Convert the camera geometry of an input cube(s) to match a specific master.

usage: matchmap.py [-h] -d DATA [DATA ...] [-m MAP] -o OUTPUT

optional arguments:
  -h, --help            show this help message and exit
  -d DATA [DATA ...], --data DATA [DATA ...]
                        Source cube file
  -m MAP, --map MAP     Input map
  -o OUTPUT, --output OUTPUT
                        Output Directory

Examples:

Process a directory full of cubes to match a single camera geometry.

mkdir recammed
matchmap.py -d *cub -m 1173J2-002_Vg2_CALLISTO_CLEAR_1979-07-08_14.06.23.cub -o recammed

About

Scripts for Cassini, Juno, Voyager, & Galileo Imagery Processing

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 99.3%
  • Shell 0.7%