The moon during these days is High, actually this weekend there was a super Moon on our skies so we can’t do deep sky astrophotography because of the moon shine that enlightes all the atmosphere. We wanted to use these days to do some elaborations but in order to be ready to achieve interesting results, we had to review our calibration library. The calibration is the way we have in order to obtain normalized data from our system. Normalized means in theory that all the data are purified from the acquiring system artefacts introduced by our acquiring system. This is very important because it’s very useful to have all the images elaborated in the same way which permit also the scientists to use our calibrated data available as usual in RAW.
As the standard for astrophotography, our calibration dataset include bias, dark and flat subtraction. Let’s go in depth on this method:
The factors that we usually take in account are:
- Thermal noise of the sensor
- Electronic noise of the acquiring and signal conditioning systems
- Defects on the optical systems
The thermal noise of the sensor, is due to the fact that the camera sensor is able to convert photons into electrons thanks to the photoelectric effect. After some hours while the sensor is working, without a proper cooling system (such as Peltier cells) the sensor can reach high temperatures and this leads to inaccurate readings by the sensor. Thermal noise is perhaps the most relevant contribution to the whole amount of artefacts in one acquired image. The common practice in astrophotography to try mitigate this effects is by compensating through Dark frames subtraction.
A Dark frame is just an image taken with the same exposure time of the sky images but with the telescope tapped. This is intended to isolate the “only” contribution of the thermal noise which can then be subtracted to all the sky images using an image processor software. However as we will see, the dark frame includes also some informations about the electronic noise. The most important thing is that dark frame subtraction can only mitigate the non-random component of the thermal noise and the hot pixels. The random component of the thermal noise is mitigated by stacking an average of identical sky light exposures but we will analyze that in a post dedicated to SNR increasing techniques.
The electronic noise is often subtracted by using Bias frames. The electronic noise is due to the elesctronic readout board inside the acquiring system. Usually to let the sensor work properly, a signal conditioning is applied after the sensor acquisition and most of the electronic noise came from the signal conditioning (such as amplifiers) and the electronic reading of the single pixel value. The acquiring electronics in fact, scans the sensor to acquire the pixel value on a single row/column on the sensor, this procedure know as readout, can address some artefacts. In order to isolate the electronic noise artifacts, it’s common practice to acquire bias frames taking some fast shots in a row with the fastest exposure settings (in order to avoid thermal noise) and with the telescope tapped. While electronic noise is already present in dark frames (that makes biases frame far from mandatory) it’s common practice to isolate anyway the electronic signal to have a better control during the processing.
Optical System Calibration
The optical system may be affected by a lot of defects, the most common are vignetting and sensor surface cleaning issues. In our case, our field of view is slight vignette and this is due to the large format of the sensor (currently we’re operating with a full 35mm frame sensor) and even if we’re using a big 2inch focuser tube, the FOV is slight vignetted. Also the cleaning of our sensor is nor perfect, actually it is since maybe 6 months that we don’t clean the surface of the sensor and since it is often exposed for several hours to the air (during the exposures) it captures dust on its surface. Taking flat field frames we are able to take the optical defects in account. It’s common practice to isolate this artifacts by tacking pictures using the same optical setup (same telescope, camera, filters and camera angle) to a bright white screen. The result is a flat field image that if the optical system had been perfect it would be perfectly white but instead, the optical defects will result in dark areas. Mathematically, the flat field are divided from the signal.
Don’t forget to calibrate your data!
That’s a brief overview on calibrating procedures used in our observatory and in general in several astrophotography facilities.
I suggest all the upcoming astronomers to preserve their calibrated data in order to have a storage of data which should be available in the future for image comparison and to have a standard calibrated image useful to be shared with other people.