Material for VIMS-IR uncertainties estimations

About

The material available here is a supplementary material related to the paper Photometric Uncertainties of Cassini VIMS-IR Instrument submitted to Icarus by Cordier, D. , Seignovert, B. , Le Mouélic, S. and Sotin, C. .

The main goal of this work is to estimate photometric uncertainties remaining in VIMS-IR cubes after calibration using USGS ISIS3 tool, the employed method relies on the definition of a collection of 3x3 pixels picked-up in each considered cube as it is summarized in the following sketch:

within each individual "block" or "box", for a given spectral channel, the variability of photometric quantities is computed. The mean relative variability gives an approximation of uncertainties.

Setup

The scripts present in this repository requires a recent python version (≥ 3.7). You can install the required dependency with the following steps:

python -m venv venv
source venv/bin/activate

python -m pip install -r requirements.txt

Known issues:

Fix HDF5 install error on macOS

On macOS you may need to install hdf5 and c-blosc libraries manually before running pip install:

brew install hdf5 c-blosc
export HDF5_DIR=/opt/homebrew/opt/hdf5
export BLOSC_DIR=/opt/homebrew/opt/c-blosc
python -m pip install -r requirements.txt

---

In addition, the provided Python modules VIMSU_1.py, VIMSU_2.py and VIMS_uncertainties.py should be available in the current directory.

Data

• VIMSuncert_cubes_list.csv: CSV file containing the full list of the 149 high spatial resolution VIMS cubes analyzed. These cubes have a spatial resolution better than 35km/pixel.
• VIMS_CALCUBES/: the directory containing all the 149 calibrated cube involved in the study. Using the list in the mentioned CSV file, all these cubes (total size: 313MB) may be automatically downloaded from Nantes University repository (https://vims.univ-nantes.fr/) with the PyVIMS tool.
• since the analysis relies basically on a Monte-Carlo algorithm, we provide precisely used in our study under the form of 2 HDF5 files available in the directory ANALYSIS_HDF5/:
  • stoDFrame_CubeData_NEW.hdf5 (48 KB) containing global features of employed cubes.
  • stoDFrame_PavData_NEW.hdf5 (79 MB) containing the data of 3x3 pixels blocks picked-up in cubes. this way, any colleague, can potentially reproduce exactly the work we have done. Each of these compressed file contains a Pandas DataFrame which can be read and use as it in provided Jupyter notebooks. Of course, the interested reader can generate his/her own HF5 file.

Tools

The global analysis process is split into 2 steps:

1. the random choice of 3x3 boxes in the cubes and the record of extracted data, stored in Pandas DataFrames, in HDF5 files. This step can be accomplished with the Python Jupyter notebook:

   • VIMS-IR_uncert_Part_ONE.ipynb

   An example of spatial distribution of 3x3 boxes is reported in the following map of Titan surface, the colored encoding represents the surface density of chosen boxes:

   

   For memory, the Huygens landing site has been marked with a red square, while Selk crater (planned to be explored by Dragonfly) is tagged with a yellow square. Titan surface mosaic a credit: NASA/JPL-Caltech/Univ. Arizona a https://photojournal.jpl.nasa.gov/catalog/PIA22770

2. the data analysis strictly speaking, which can be done with the second Python Jupyter notebook:

   • VIMS-IR_uncert_Part_TWO.ipynb

Procedure

1. check if all the required Python modules are propertly installed on the machine.

2. install all the present package, with or without the VIMS cubes available here (they can be downloaded automacically by the notebooks).

3. use the first notebook VIMS-IR_uncert_Part_ONE.ipynb in which the name of the directory containing the VIMS cubes will be asked, this directory has to be available on your hard drive, even if it is empty. In this notebook, the user has to specify:

   • the name of the directory containing the cubes
   • for processed cubes, the fraction (0.10 is a reasonable value) of pixels that will be chosen as the "central pixel" of considered 3x3 boxes.
   • the names of the HDF5 compressed files containing the data:
     • one with global features of cubes (exposure times, etc)
     • on with data of 3x3 boxes.

   As output, the user will get:

   • pictures of cube (at 2.03 microns) with chosen pixels identified by a red dot, these pictures are shown in the notebook itself, a copy of each one are saved in the directory TEST_Plot_Distrib/, if TEST/ is the name of the directory containing the cubes. These plots, similar to that one:

     

     will be also shown in the notebook.

   • two HDF5 compressed file, mentioned previously.

4. use the second notebook VIMS-IR_uncert_Part_TWO.ipynb, this notebook is dedicated to data analysis, here an example of that could be obtained with this tool:

   

License

The source codes in this repository (*.py and *.ipynb) are provided under a open-source GPLv3 license. The output results (ANALYSYS_HDF5/ and fig/) are distributed under a CC-BY-SA license. The VIMS_CALCUBES/ were downloaded on the VIMS Data Portal.