Sony A7S all-sky camera

Sony A7S all-sky camera#

Where is the data?

Quicklook images at birkeland, archive at NIRD

Who?

Dag (PI) and Mikko

Where?

Module 3

Main operational task

Starting and stopping of the imaging

Warning

The first Sony images up to and including 14 Nov 2015 are out of focus

Warning

The date and time in the EXIF-data in each JPEG-image are incorrect. The correct date and time are embedded in the filename.

Data description#

The Sony camera stores colour all-sky images in JPEG-format. The image cadence is roughly one image every 10-12 seconds. Also, there are quicklook images for realtime webpage: these have a small label indicating date and time of the image.

The summary plots are keograms covering the dark time of each day.

Note

The filenaming convention is a bit confusing, but the raw files are LYR-Sony-ddmmyy-HHMMSS.jpg where dd and mm are the day and month, yy are the last two digits of the year and HHMMSS are the hours, minutes and seconds of the UT time (start of exposure). The keograms are named slightly differently with a full four digit year LYR-Sony-yyyymmdd.png.

The data from Sony is archived to NIRD.

Star calibration#

Examples and source code for doing a simple star calibration can be found at github

Computer setup#

Control computer

  • Ubuntu 14.04

  • Use ssh for remote access

Note

When starting the imaging software, wait for a password prompt! This may take a while… Also, you will need to use a GUI which – when controlling the software remotely – requires all sorts of ssh-tunneling via VNC etc. (which does not seem to work well from Windows…)

Configuration#

Note

The Sony camera body does not include a GPS and the current imaging software does not check nor update the date and time in the camera. This results in incorrect EXIF-data in the images. One should set the date and time to match the season when powering up the camera at the start of the season.

The configuration of the python program is done in the .pysces_asi/settings.txt file.

As of December, 2020, the following settings for exposure time etc. are in use:

Mode

Min time between images

ISO

Shutter speed

Notes

Nighttime

10s

16000

4 s

Day Time1

30s

100

1/1000s

Not actually used?

Day Time2

30s

100

1/2000s

Not actually used?

Twilight1

30s

100

1/8s

Twilight2

30s

200

1s

Twilight3

30s

2000

1.3s

The mode is selected based on the location of the Sun and the Moon in the sky, but not on the phase of the Moon. One should also note that in LYR the Moon is always less than 40 degrees above the horizon during wintertime… Here is the selection logic:

SUN_ANGLE < 0 and SUN_ANGLE > -3 and MOON_ANGLE < 40 = "Twilight1"
SUN_ANGLE < -3 and SUN_ANGLE > -6 and MOON_ANGLE < 40 = "Twilight2"
SUN_ANGLE < -6 and SUN_ANGLE > -9 and MOON_ANGLE < 40 = "Twilight3"
SUN_ANGLE < -9 and MOON_ANGLE < 40 = "Night Time"

There is a backup system disk image in KHO-WD4

Software#

The paskil software was originally developed by Nial Peters (2009). For details, see his Master’s thesis http://urn.nb.no/URN:NBN:no-22683

There are also pysces_asi and paskil python libraries being used. There are copies in NIRD and in github (paalge)

  • Instructions for starting and stopping of the camera operations are as the desktop background image. Originals can be found in khodata:\Doc\Instrument instruction sheets\Sony_a7s

  • The software will ask for password for various network drives. It may take some time before the dialog window appears (i.e. several minutes)

  • Data management routines are operated via crontab

Drawbacks in the current implementation#

The current software and setup is a result of incremental additions to the original software. Especially the data management routines are a bit convoluted and reflect various changes in KHO infrastructure (local disk, NAS-box, NIRD). Also, pysces_asi writes huge log files, which is probably a left-over from earlier debugging efforts.

The scheduler appears to be free-running rather than tighly synchronised to system time. This makes the timestamps in filenames inconsistent, which then requires additional care when working with data. The directory structure for data and how the routines pack images into tar-files for archival is somewhat peculiar.

For example, here is what comes out when unpacking archived tar-files 20200107.tar.gz (note missing directories for year and month):

06/
06/Images/
06/Images/LYR-Sony-070120_133004.jpg
06/Images/LYR-Sony-070120_133015.jpg
06/Images/LYR-Sony-070120_133027.jpg
06/Images/LYR-Sony-070120_133038.jpg
06/Images/LYR-Sony-070120_133050.jpg
06/Images/LYR-Sony-070120_133102.jpg
06/Images/LYR-Sony-070120_133113.jpg
06/Images/LYR-Sony-070120_133125.jpg
06/Images/LYR-Sony-070120_133137.jpg
06/Images/LYR-Sony-070120_133148.jpg
06/Images/LYR-Sony-070120_133200.jpg

One more practical issue is that the date format in the filename is day-month-(incomplete)-year . Perhaps, in future we might have time to re-write parts of the code to result in more user-friendly archived data packages. Also, the image capture times could be consistently fixed to, for example, every 10 seconds during nighttime:

2020/01/LYR-Sony-20200107_133000.jpg
2020/01/LYR-Sony-20200107_133010.jpg
2020/01/LYR-Sony-20200107_133020.jpg
2020/01/LYR-Sony-20200107_133030.jpg
2020/01/LYR-Sony-20200107_133040.jpg
2020/01/LYR-Sony-20200107_133050.jpg
2020/01/LYR-Sony-20200107_133100.jpg
Contents