panoptes-timegen

About the Timelapse Generator

The Timelapse Generator is capable of outputting Full HD video (1920 x 1080) or the original frame size (5208 x 3476) to an mp4 video file. The codec fourcc tags used are avc1 and mp4v. The first choice is avc1 and on failing to use this mp4v is used.

The Timelapse generator takes a directory containing FITS files, either *.fits or *.fz. Please take care to avoid putting *.fz and their decompressed *.fits in the same directory to avoid duplicates in the video as the script doesn't differentiate between them.

Each FITS file from Project PANOPTES is usually of size (5208 x 3476) and is usally single channel. The timelapse generator takes these type of files, and applies the following operations on them:

• Debayering to get RGB image. Bayer patter used: 'RGGB'. Remember that astronomy format is to save zeroth pixel at bottom left corner of the image array
• Stretching to modify the contrast of the image.
• Sub-image extraction to get only a part of an image. Note that the script rotates any sub-image to ensure that the video is always in landscape mode.
• Saving the sub-images in .jpg format to a directory called temp_timelapse in the same directory as the time_gen.py script.
• Generating the video file from the .jpg files stored in the temp_timelapse directory.

After successful generation of the video file, the temp_timelapse directory is cleared.

Setting up the environment

Windows (Tested on Windows 10)

Download and extract the github repository. Say its at \panoptes-timegen-master. Open a new CMD window as administrator and do the following:

Make sure you already have python 3 installed and accessible from the command line as python.

• Step 1 :

  • In the shell, run pip3 install venv
  • python -m venv pan-tg This results in a new python virtual environment named pan-tg at \panoptes-timegen-master\pan-tg

• Step 2:

  • Activate the virtual environment by running the shell command: pan-tg\Scripts\activate from the panoptes-timegen-master directory.
  • Install the required python packages by running pip3 install -r requirements.txt after activating the virtual environment

Linux (Tested on Ubuntu 18.04)

Download and extract the github repository. Say its at /panoptes-timegen-master. Open a new terminal and navigate to this directory.

If you already have pip3 and venv installed, skip Step 1.

• Step 1 : In the terminal do:

  • $ sudo add-apt-repository universe
  • $ sudo apt-get install python3-pip python3-venv

• Step 2 : Now start installing the required python packages by activating the virtual environment.

  • $ python3 -m venv pan-tg
    
  • $ source pan-tg/bin/activate
  • (pan-tg) $ pip3 install -r requirements.txt

Here pan-tg refers to the virtual environment where we run the code. Creating a virtual environment is recommended for this project.

That's it. The environment is setup now.

Running the Timelapse Generator:

If you already have the FITS files generated by PANOPTES units, directly go to Step 2

• Step 1:

  • Get the FITS files from Panoptes Observations Explorer by choosing the sequence from the list and downloading the text file containing the urls.
  • You can download the FITS files using wget or curl or you can use the provided download.py script provided for convenience.
  • USAGE:
    (pan-tg)$ python3 download.py --help to see the help text Linux
    (pan-tg)$ download.py --help to see the help text Windows
    (pan-tg)$ python3 download.py -i <path_to_text_file> -o <path_to_dir> Linux
    (pan-tg)$ download.py -i <path_to_text_file> -o <path_to_dir> Windows

• Step 2:

  • Activate the virtual environment that has been setup as described in the OS specific instructions above.
  • The timelapse generator is called as follows:
    (pan-tg)$ time_gen.py -i <path_to_fits_files> -o <output_path>
(pan-tg)$ time_gen.py --help to see all options.
    (pan-tg)$ time_gen.py -i <path_to_fits_files> is the simplest to use.
    (pan-tg)$ time_gen.py -i <path_to_fits_files> --full_hd to force output video to be 1920 x 1080p
    (pan-tg)$ time_gen.py -i <path_to_fits_files> --m 1 --n 2 --cell 0 1 to specify that the image is to be divided into 1 Row (m) and 2 Columns (n) and to get the sub-image specified by cell (0,1) i.e 0th row index and 1st column index.
  • On Ubuntu, you will have to prefix the commands with python3 to get them to work. The shebangs were removed because they broke Windows support. Example:
    (pan-tg)$ python3 time_gen.py -i <path_to_fits_files> -o <output_path>
(pan-tg)$ python3 time_gen.py --help to see all options.
    (pan-tg)$ python3 time_gen.py -i <path_to_fits_files> is the simplest to use.
    (pan-tg)$ python3 time_gen.py -i <path_to_fits_files> --full_hd to force output video to be 1920 x 1080p
    (pan-tg)$ python3 time_gen.py -i <path_to_fits_files> --m 1 --n 2 --cell 0 1 to specify that the image is to be divided into 1 Row (m) and 2 Columns (n) and to get the sub-image specified by cell (0,1) i.e 0th row index and 1st column index.

Performance:

The timelapse generator takes around 20 seconds to process each FITS file. The video generation takes additional time and is faster if the --full_hd flag is set.

For debugging purposes, logging has been setup. The logs.log file contains the log from the most recent run of the timelapse generator.

Troubleshooting Help:

• Check if you are running from the right virtual environment if you have configured it.
• Check the version of the Python that the script is running from. The scripts are written to work with Python 3.
• If Windows complains that it couldn't find libopenh264-*.dll you can safely ignore it. We're not using libopenh264 components here.
• If using mp4v encoding, the output file may not play with the default codecs present on the system. It is recommended that you try opening them with vlc media player.
• The output timelapses are currently tinted green. This issue is being considered and we're exploring ways to correct it.