| Are Starlink satellites an issue? | |
|---|---|
| Starlink is a big issue for me | |
| Starlink is a non-issue | |
| Starlink is an issue, but can be handled in post-processing | |
| Login to vote and view results. |
11.01
#...
·
 |
|---|
|
There was a lot of fuss about SpaceX's starlink network of satellites causing problems for astronomy/astrophotography a while ago. Now that there are thousands of starlink satellites up there, I wondered what people think? I see them in many of my subs, but I personally don't see a big issue as the rejection algorithms seem to work quite well. |
7.75
#...
·
·
4
likes
|
|---|
|
I believe we have to distinguish between current situation and future prospects. Currently it can get handled easily in post processing. I'm really concerned about what the sky will look like in the future. In April 2021 a total of 7389 satellites in space have been counted by United Nations Officer for Outer Space Affairs (Source: https://www.geospatialworld.net/blogs/how-many-satellites-are-orbiting-the-earth-in-2021). Starlink alone intends to add up to 42.000 additional ones. Other companies have similar plans although with more conservative number of satellites (source: Starlink alternatives: 3 SpaceX rivals you need to know (inverse.com)). At least Starlink seems to listen somewhat to concerns of astronomers. Clear skies Wolfgang Edit: replaced "United States" typo by "United Nations" |
3.34
#...
·
|
|---|
| I think SpaceX get a bad rap for starlink because they are a single entity putting up lots of satellites. It seems that every single trail caught by astrophotographers is attributed to Starlink. The fact is that, yes they are putting up a lot of satellites to benefit the entire planet, but they are only joining the list of trail makers. Other satellites, meteors, the space station, airplanes..... the list goes on. |
10.79
#...
·
·
7
likes
|
|---|
Ryan Jones: According to this article, SpaceX passed the 2500 satellite mark this May: https://spaceflightnow.com/2022/05/13/spacex-passes-2500-satellites-launched-for-companys-starlink-network/ Based on this data, there were about 5000 active satellites in Orbit in 2021: https://www.statista.com/statistics/897719/number-of-active-satellites-by-year/ Give the fact that quite some non-SpaceX satellites in orbit are on geostationary orbits (about 400 in 2015, source, which was about a third of sats at that time), the likelihood is high that any observed LEO object is a Starlink satellite. Probability increasing. At the moment it's only SpaceX but other companies have announced plans of super-constellations with numbers of satellites of the same order of magnitude. Personally, I consider our night sky as part of nature like the rest of the universe and I prefer asking this question: Our actions have an effect. What's the effect and is it worth it? Regarding these super-constellations, my answer is: No. It's essentially littering space. For AP: at the moment, math helps. For the future: I'm tentative to say that math helps but I wouldn't be on it that it fully solves all issues. Besides: I am curios how many of you ever look up to the sky? If you do, then say good bye to the night sky because many moving objects in the sky will prevent you from focusing on the fixed objects as we tend to follow moving objects with our eyes. That's my two cents about it. BJörn |
|
3.34
#...
·
|
|---|
| Let’s call LEO 2000km in altitude and the radius of earth an even 6,300 km. So the surface area that these satellites roam around on is 860,000,000km2. Let’s throw 40,000 satellites up there. So about 21,000 square km between each one..... there’s a fair bit of room up there..... yes I get the need to be responsible with what we put up there and sensible with how we bring it down but let’s not get carried away and stop progress. |
|
10.79
#...
·
·
4
likes
|
|---|
Ryan Jones: 21,000 sq.km may sound like a lot but that means only an average satellite distance of about 160km (if they are in the same plane). Also, I don‘t think that one should take the Kessler syndrome lightly. Regarding progress: I see that progress is a term lightly used to justify anything. What does progress mean? From experience, it rather seems to me that mankind repeats the same mistakes all the time: 1. Let’s act and see the consequences later. 2. Progress: most of the times it results into „new“ and „different“ but in essence essentially the same. Björn |
|
7.42
#...
·
|
|---|
Ryan Jones: I think the maths is wrong. It is the apparent size, i.e., magnitude, times orbital frequency that matter here. In certain locations the likelihood of repeated passes is very high thus reducing the effectiveness of any rejection algorithm. |
|
7.75
#...
·
·
1
like
|
|---|
|
I assume we won't get agreement on whether Starlink really benefits the entire planet or just a limited number of privileged people + Elon Musk and whether it's really progress. This is not the forum to discuss this anyway. Take your current percentage of images with satellite trails and multiply it by five to ten as a rough estimate what you have to expect. Then decide if it bothers you or not. If professional astronomers - most likely with more elaborate methods to deal with satellite trails at their finger tips - raise their concerns about the current development, we should take it serious in my humble opinion. If it starts hurting our earth based astronomy it gets in the way of science and hinders progress. On a side note: Starlink currently operates at 550km and has plans for >20k satellites between 330 and 400 km height. So your calculation is off by a factor of 4 to 6 in height or 16 to 36 in area. I would calculate like this: the sky spans 360 degrees right ascension and 180 degrees declination leading to 64.800 1 degree x 1 degree sections in the sky around us. With 42k Starlink satellites + 7k existing satellites + something like 10k satellites by other players coming up we're approaching on average 1 satellite per section. Of course there are preferred positions for satellites so it's a fairly inaccurate calculation. On the other hand it's bad news for these preferred areas of the sky. Admittedly there are more severe problems to tackle on this planet ... Clear skies Wolfgang |
4.69
#...
·
|
|---|
| I don't really see the problem. If the occasional sub has a white line on it, I just leave it in as it will be removed in integration anyway (since those pixels are not lit up in any of the other subs) |
|
7.42
#...
·
·
1
like
|
|---|
Stuart Taylor: Except it doesn't actually work that way. |
|
4.69
#...
·
|
|---|
andrea tasselli:Stuart Taylor: Sure it does. "Satellite trails are removed by the sigma clipping algorithm: Almost every modern astronomical post-processing program has a rejection process (sometimes referred to as sigma-reject) to remove unwanted signals, though the exact sequence will depend on which program you use.  Typical rejection settings for stacking in PixInsight.The way this process works is that, while averaging all of the pixels in a series of, say, 10 images, the program mathematically calculates which pixels fall far away from the mean value because they're much brighter (or much fainter) compared to the same pixels in other frames. The algorithm then discards those out-of-range pixel values so they don’t affect the final image. This process easily removes satellites, airplanes, and UFOs from your final stacked image.Some programs, including PixInsight, even create a "rejection map" that displays only the outlier pixels that were not used. If you take a look at this image, you’ll see cosmic ray hits, a single slight tracking error or bump, the occasional airplane, and (increasingly) many satellite trails." |
2.11
#...
·
·
3
likes
|
|---|
Stuart Taylor: "Removed" is subjective, and this isn't an ideal solution. WSC, given enough data, and normalized data, can actually do a respectable job of handling satellite trails. It will still typically leave some signal behind for bright trails (which StarLink tends to leave) which will add contamination to images where you're really pushing faint signal. The other problem with cranking up WSC settings is that it comes with a broader sacrifice to signal, which also isn't ideal. Generalized Extreme Studentized Deviate (ESD) rejection, with a suitable dataset, can be more effective in some regards, as it is much kinder to other forms of signal. But it still has limitations. One of which is that it depends on a certain subset of frames representing deviation, and as pixels feature more and more outliers, relative to the number of frames captured, this cleaner approach to rejection can be complicated. Right now Star Link really isn't so much an issue because it tends to be pretty limited in scope—limited in terms of actual impact on imaging. Sometimes, perhaps, giving a nice big surprise on a particular session. But the broader concern for imaging, I'd argue, is that StarLink is only the beginning. Only the beginning of what SpaceX plans to put up there. And only the beginning in terms of similar satellite networks other competing interests plan to put up there. The potential of the issue, setting aside other scientific concerns, is a network which is far, far more expansive and intrusive than what we are dealing with right now. |
|
7.42
#...
·
·
2
likes
|
|---|
| On top of what already said by James there is also the potential (but in reality quite real if you image long enough) to have the same satellites trail happening close enough to each other. If you think an entire train of these, even ESD rejection would have issues, never mind the sacrifice in SNR you have. I forsee signficant problem going forward. If you're buying time on a remote system then chances are there that you won't be able to use rejection algorithms as affectively as at home (long integrations with CCDs vs. short ones with CMOS) or not at all. It is a situation I witness far too often. |
6.77
#...
·
·
1
like
|
|---|
|
Right now I think it's still manageable. But.. yea, the future might be very different. There is a big difference between a few thousand and tens of thousands of satellites up there.. 2500 starlinks today, 42.000 eventually, of just SpaceX, but other companies are planning on doing similar stuff. So eventually it will not just be 1 line in the occasional sub, but might turn into a line in each sub. So I can say that I'm a little 'worried" about the impact this eventually will have. Not just on photography, also the visible night sky, and of course the increasing possibility of the doom scenario where satellites start hitting each other in a chain reaction. (Given how often it is in the news already that starlinks are getting too close to the ISS and such) But, at the same time I guess, stacking software might get better and better to deal with it. Mainly because they are predictable paths, where software might be able to clone out the trails better and better once it knows times/dates/locations/etc to compare that to the image instead of just comparing pixels as it does now. |
|
7.42
#...
·
|
|---|
So eventually it will not just be 1 line in the occasional sub, but might turn into a line in each sub. Depending on location and time is already much more than that. I get 1 or 2 trails every 4-5 frame on occasion. In few cases I get 3 trails in one frames. Because they are brighter than most of everything else I have to screen suspected ones for reference frame during normalization. |
|
4.69
#...
·
|
|---|
|
I don't see them often in my subs. Is it possible that I am just in a lucky location? Or do these satellites basically cover everywhere equally? EDIT: actually, I think I know the answer to that one. If they're providing internet access they would need to be evenly covering the earth |
|
7.42
#...
·
|
|---|
|
It depends on how often you image, where the scope is pointing at and finally where are you. I guess some locations far to the north may see less than more southernly locations but this is my guess. And to show how potentially significant a alrge increase in trails is, here is a stack of 18 180s subs, with a rather strong rejection filter applied to. There are at least 3 visible traces left, the ones with highest magnitude:  |
|
10.79
#...
·
·
2
likes
|
|---|
|
Just found this news: "New center to coordinate work to mitigate effect of satellite constellations on astronomy". (w.r.t. my previous post item 1: q.e.d.) For those who are living closer to the poles, during the respective winters, the satellites on lower orbits will enter Earth's shadow quickly and so won't be visible. In the transition and summer periods, the satellites will be longer or even be permanently visible. From own experience, when I star imaging during spring or fall, the early images and late images (early morning) contain most trails (and virtually each sub). Based on this wikipedia entry (section "Constellation design and status") we can make some back on the envelope estimations: Given 72 equidistant orbital planes, your FOV will cross an orbital plane every 10 minutes. The orbital period of a satellite is approximately 100 minutes. With 22 sats per orbital plane, every 4,5 minutes, a satellite will cross your FOV. Extrapolating to a full constellation, a satellite on each sub is basically guaranteed. Of course, the question is, if sun is illuminating them (see above). |
|
11.01
#...
·
|
|---|
andrea tasselli: I was imaging M106 the other night and saw satellite trails every 3 or 4 images. I live in England, quite close to London |
9.03
#...
·
|
|---|
|
Is it true that they have mitigated the problem by reducing reflective surfaces on the satellites? This says not really, despite the press releases: https://physicsworld.com/a/dark-coated-starlink-satellites-are-better-but-not-perfect-say-astronomers/ |
|
10.79
#...
·
|
|---|
Is it true that they have mitigated the problem by reducing reflective surfaces on the satellites? This says not really, despite the press releases: According to a recent online article, current satellites don’t contain the brightness reduction as it affects the satellites. (Only found the article in German but I guess some online translation will help you out) In addition, the same website has an article about the impact on different regions. In essence regions at about 50 North and South are impacted most heavily. |
3.34
#...
·
·
1
like
|
|---|
I believe we have to distinguish between current situation and future prospects. Currently it can get handled easily in post processing. Share your concerns, Wolfgang. How long will it take until each and every telecom company realizes they need sats for 6G, 7G, 8G.... it will be a huge problem in the near future. At the same time, we need R&D in making sats less bright and an astronomer community that speaks up, is actively and effectively engaged in multilateral settings as well as with companies such as StarLink and others. |
|
7.75
#...
·
·
2
likes
|
|---|
I suggest to name the StarLink satellites "Muskitoes"  . They are as frequent and as annoying as their living counterparts.Clear skies Wolfgang |
|
10.79
#...
·
·
3
likes
|
|---|
|
Here’s a recent and interesting video from Palomar‘s YouTube channel: https://youtu.be/477TvMF7WEA |
7.61
#...
·
·
1
like
|
|---|
Here’s a recent and interesting video from Palomar‘s YouTube channel: That was very scary...... I was recently imaging in Ursa Major (NGC 3718 from lat 44 N) and on average had a satellite cross my fov every 7 min. Within the 2 hours of imaging time, I also had 4 subs with multiple trails, so every 30 min on average. When just SpaceX has reached its initial target, it won't be a question of how many subs have trails, since they all will, it'll be how many have 2, 3, or 4 trails..... The real problem is that most people that are in any position to effect something like this, by virtue of living in urban, and many suburban, environments don't have a night sky to lose. They, in fact, don't even know what they've already lost. To them, Starlink (and the others to follow) is just another internet option to choose from, and even if just a little cheaper, or faster, or more dependable, it's all gain and no cost...... Scott |
