Correct setup of planetary imaging cameras

Do camera parameters really matter?

In planetary imaging, it is often said, high frame rates with the benefit of capturing as many frames as possible is the one and only key to succesful lucky imaging. In this article, we want to dive a bit deeper and evaluate the effect of different camera settings on the final image result. For that series of experiments, Mars was captured in several consecutive runs at the same night without any change of focus or image scale within a time frame of less than ten minutes from the preconditioned dome at Juniper Hill Observatory, assuring, that both the dome climate and the imaging train were in complete thermal equilibrium. Optical setup The optical setup was a 190mm (7.5“) Maksutov Newtonian combined with a Televue 5x powermate. In addition, a Baader IR/UV block filter was used to limit the captured spectrum to VIS only, avoiding blur due to miisfocused NIR. As imaging camera, a ZWO ASI 224 MC color camera was used to avoid any influence caused by derotation processes necessary in LRGB-imaging sequences using a monochrome camera. The effective focal length of this setup was about 6600mm or f/35 for the given aperture. The images were later stacked with a drizzle factor of 1.5x to increase the visibility of small albedo structures. Note, that the apparent diameter of Mars during capture was 14.2 arcsecs. Capture interval was Jan-03-2023 from 18:10 to 18:16 UTC. Test scenario The test scenario should answer one basic question: Is it better, to go with a minimum exposure time and highest possible fps at the cost of high gain, higher noise and lower dynamic range or go the oppsite way with low gain for a maximum dynamic range with lowest noise, but high exposure time and low fps. Let‘s first look at the camera charactersitics, so that we can later understand the third capture scenario of this test:

Source: https://astronomy-imaging-camera.com/

We have chosen the following scenarios for high and low gain: 1 . High gain: Gain = 325, exposure time = 6.675msec, fps = 149, total frames = 18000 (120sec) 2 . Low Gain: Gain = 153, exposure time = 44.51msec, fps = 22, total frames = 2700 (120sec) As the camera is - as is typical for modern CMOS planetary imagers - operated in High Speed mode, the ADC runs on 10bit resolution instead of 12bit. If you check the blue curve above (DR stops), you can see, that with the high gain settings of Gain=325, you have a dynamic range of 9bits, whereas in our low gain scenario, at Gain=153, you haver nearly 12bits of dynamic range. So in low gain, your noise is below the sampling threshold, whereas in the high gain scenario, the noise affects the 2 LSBs of the sampling process. An optimum gain could be around a value of 250 with a DR of 10bits, where the SNR perfectly matches the ADC‘s resolution. So we have a third scenario: 3. DR10 gain: Gain = 250, exposure time = 16.13msec, fps = 61, total frames = 7500 (120sec) Let‘s first have a look at the quality graphs of the three scenarios: 1 . High Gain: 2. Low Gain: What can clearly be seen, is, that in the high gain scenario you are facing a quite steep quality graph with lots of frames with poor quality, but with much more frame than in the low gain scenario. The low gain scenario shows a relatively flat quality graph with lots of frames being about a certain quality threshold: this is mainly due to the fact, that the associated long exposure times average out high frequency seeing effects, resulting in less sharp, but more consistent frames. Let‘s now see the compromise scenario, whre SNR is matched to the ADC resolution: 3. 10DR Gain: As to expect, the quality graph lies somewhere in between the scenarios 1 and 2. As a next step, each of the three scenarios has been stacked with a frame selection rate of 15%, and then sharpened using Registax wavelets. The result can be seen in the following image. Klick for a 100% view: The clear winner here is the 10DR Gain scenario, with the Gain adjusted to match the 10bit resolution of the ADC. It shows at least the fine details of the High Gain scenario, but with the smoothness and clarity of the Low Gain scenario, where the latter is missing the finest detail due to long exposure times. To be honest, the differences are quite subtle, so neither imaging scenario is a real show stopper and it might make sense, to chose the parameters in a way to best meet the circumstances (seeing, focla length, pixel size etc.) of your imaging setup. But in general, reducing the gain to a value of 10bits DR drastically reduces disk space and processing time and yields cleaner and more detailled results than sampling too fast using high gain values. What definitely should be avoided, is to shorten exposure times below the frame rate that your camera can achieve with the selcted ROI. For example, if your camera supports 100fps at an ROI of 800x600 pixels, it makes no sense to push your exposure below 10msec: you do not get more frames but you get considerably more noise and lower dynamic range. So take 1/max{fps} as your minimum exposure time and try to limit your gain to a value giving you 10 stops of dynamic range. If your planetary camera supports high conversion gain mode (this is, where the readout noise drops at a certain gain value) NEVER do planetary imaging with gain values below the HCG mode. With that points considered, one can achieve a maximum detail level in planetary imaging without generating excessive amounts of data.