Thayne McCombs and Dr. Joseph Ward Moody, Physics and Astronomy

Introduction

The ROVOR observatory is a robotic, remotely controlled telescope in Delta, Utah. Since it is remotely controlled and semi-automated, it produces a very large amount of raw data. Unfortunately it took much longer to process and analyze the data than to acquire it, and, therefore, we had a backlog of over two years of data that was waiting to be processed. This delay in analyzing data was detrimental to monitoring projects because without the analysis of recent data it is dicult to determine the most interesting targets to observe.

Producing a graph of how an object’s luminosity changes with time from raw images requires a number of steps, including zero correction, dark correction, at correction, astrometric solution, and photometry. Much of this process can be automated. However, traditionally, each step is performed manually using a software package called IRAF which is somewhat clumsy to automate, and time consuming to use. The ROVOR project would greatly bene t from a more ecient method of processing data. My objective was to develop software which would automate as much of the data reduction process as possible, and provide an easy to use interface for the parts which required human intervention. The software should also be able to run on the observatory site, and eventually be triggered by software that controls the telescope itself.

Methodology

Since the reduction software needed to be located on the ROVOR site, and accessed from BYU campus, it seemed natural to use a web interface to interact with the software. However, the local system should also be able to access the reduction routines directly. As such, we decided to separate the logic related to the reduction into an independent library, which the web interface utilized. However, the full power of the reduction could still be accessed by outside scripts without needing the web interface.

Python was a natural choice for the language because there were already libraries for manip- ulating astronomical images, and integrating with IRAF.We also used the Django framework to facilitate development of the web interface using Python. Zero, Dark, and Flat correction are relatively straightforward arithmetic operations on the pixels of the images, which we performed directly using the py ts and numpy libraries. For the astrometric solution, we decided on astrometry.net which automatically solves for the true position of the image. However, we needed to parse the headers of the image in order to determine an initial guess. Finally, for photometry, IRAF was the best resource available, and used its PSF tting routines for photometry. However, we were able to automate the determination of all the parameters which greatly reduce the amount of human time needed to perform that step. To integrate with the IRAF software we used PyRAF, which allows python to call IRAF routines directly.

The processed data needed to be separated based on target and lter. This information is stored in the headers of the images, and when photometry is actually performed for a night of observations, we separate the images into categories for each unique combination of lter and target. Then for each combination it produces a separate light curve.

The web interface uses a Model-View-Controller design which separates the logic for display information and forms from the logic that processes information and stores data. In addition to forms which allowed the user to manually control some aspects of the reduction, we created mechanisms for a login system and for browsing a section of the le system where images are stored.

Results

We developed a system which automates most of the data reduction process for astronomical images. Although it is optimized for the ROVOR observatory, it could be relatively easily adjusted to work for any observatory that performs photometric observations. It produces output of comparable quality to data reduced manually with signi cantly less e ort.

Discussion

The software we developed noticeably increases productivity for researches on the ROVOR team by automating a large portion of the data reduction process. In addition, the web inter- face is much simpler than the command line interface of IRAF, and therefore decreases the learning curve before a student can provide meaningful contribution to the research group. However, this software is just the rst step in fully automating the ROVOR observatory. The next project is to further automate the telescope’s operation so that only a list of targets must be supplied for each nights observations. Such a system could automatically begin the reduction process after the nights observations are complete.

Conclusion

We have developed software which automates a signi cant portion of the data reduction pipeline for the ROVOR observatory, and provides an easy-to-use interface to use it. This software could be adapted for other observatories, and can be automatically invoked by the same software that controls the telescope. It’s online interface also makes it easy to use remotely.