DSLR Photometry Tutorial / Intro to DSLR Imaging
Intro to DSLR Imaging
Before you start imaging and analyzing your own data, we suggest you use one of our sample data sets to learn the steps for obtaining a raw instrument magnitude. Links to these data are contained in the “Starting Analysis” section of the tutorial.
Most of the imaging procedures for DSLR photometry will be very familiar to those who have dabbled in astroimaging. The goal of photometry, however, is not to produce an attractive image but to accurately record the relative brightness of each star compared to other stars in the field. This accuracy is achieved first, by calibrating each image to eliminate both electronic and optical anomalies of the camera and second, by stacking multiple exposures. Stacking increases the signal (star photons) to noise (background photons) ratio and averages out the effects of atmospheric variability.
We begin our photometry session by taking a series of images (~10) called “darks”. Be sure to record all your images in "RAW" format. Darks will be used to eliminate the electronic anomalies of the camera. Darks are simply images made with the lens cap on. They record only internally generated electronic noise such as that produced by "hot" pixels. Your darks should be taken when the camera has reached the ambient outdoor temperature so you may want to set your camera outside or in the garage for 30 minutes or so before taking them.
We should note that some DSLR photometrists have suggested that darks are not necessary with exposures in the 5-second range. Also, your camera may have an automatic noise reduction feature. Using this feature makes taking darks unnecessary.
Optical aberrations are handled using another series (~10) of images called "flats". Flats are images of a blank surface such as an evenly illuminated white foam-core board. Since the target is blank, flats highlight optical aberrations like dirt on the lens or distortions such as vignetting. Computer software will process the darks and flats to eliminate these imperfections and produce calibrated star field images.
There are a number of different methods used for taking flats. Light boxes and sky flats are two popular techniques. I use a simpler approach. I take flats by imaging a white foam core board in a dimly lit room. The dim lighting allows me to move the camera around while taking a longer (~5 sec) exposure. This motion eliminates the possibility of recording surface anomalies on the board. It is essential that the board be evenly illuminated and that the optical path (lens, focus, f-stop, etc.) be exactly the same as you use while imaging stars.
With the darks and flats in hand, it’s time to go outside and image the star field. In addition to a tripod, you’ll need a cable release. A right angle viewer can also be very handy, especially when aiming near the zenith. The focal length you choose should provide a field-of-view wide enough to include not only the target star, but also an array of comparison stars. These comparison stars will be used to compute a Transformation Coefficient and determine the final "V" magnitude of the variable star.
Locating target stars can be a challenge because they will typically be too dim to see in the viewfinder. I try to identify a bright star near my target star and, knowing my field-of-view, use a star chart to determine where I should place the bright star in the viewfinder.
Once the desired stars have been located, it’s actually best to defocus the camera slightly to spread the star images over a larger number of pixels. I do this by setting the focus to "manual" and turning the focus adjustment as far as possible toward "infinity". The infinity stop is actually slightly out of focus.
Taking between ten and thirty images at maximum aperture and 800 ISO should work well. Exposure times will vary depending on the magnitude of the target star. Try a 5-second exposure as a starting point. The key in selecting your exposure time is to avoid over exposure of your stars. It’s best to keep the sensor’s maximum pixel values at less than 80% saturation. This will ensure that the camera working in its linear range. (see the linearity and saturation test below). Maximum pixel values for each star can be determined using the analysis software discussed in the next step, “Starting Analysis”.
With your star images, darks and flats recorded, you’re ready to download the images and begin processing in the next step “Starting Analysis”.
|Read More: Linearity and Saturation Test|
It’s important to know how your camera responds to light and when it saturates. When your camera saturates it no longer responds linearly to additional incoming photons and will provide erroneous readings. What do we mean by linear? Let's use the below graph to explain:
As you can see the brightness the camera detects appears to fit very well to the superimposed line until about 21 seconds. At this point the camera's response deviates from the linear trend and stars reporting fewer photons than would be expected. Everything up to this turning point is where the camera is responding linearly, everything past this point is the non-linear regime of the camera. If we were to measure a magnitude of an object in the non-linear area of our camera we would get an incorrect value because the camera would report the star as being dimmer than it actually is.
But how do we find this saturation point and avoid it. To determine it precisely would require very expensive equipment, but we can get close by a little experimentation.
Saturation can occur as a function of several parameters. The exposure time, ISO setting, and even changes in the camera's gain (if you can control it). In this document we will only discuss saturation as a function of exposure time. We encourage others to discuss test for other saturation sources in the photometry forums. To test linearity as a function of exposure time:
The brightness and time values you just determined are a reasonable first-guess for the maximum brightness your camera can record before it becomes non-linear. Keep this number in mind when you do your analysis. If the pixel brightnesses are approaching 80-90% of this value on a typical exposure you will need to treat the data with caution or even redo the measurement with a lower exposure time or ISO setting.