Derby would be nasty in the furball meta, since it’s just in the most basic sense an ARH Python 3 or 4 or 5. Too bad Israeli aircraft aren’t just good in general.
You have a humble South African Gripen with R-darter tho. Iirc it is pretty much same as Derby
I mean, it depends on how the image track is done, theres actually more then one method used for IIR tracking, however all of them are basically different methods of edge detection, with things like additional convolutional layers added on top.
Theres a few things that limit this effectiveness. Namely noise. In neural networks, of which convolutional ones are most commonly used for image detection; theres something known as an efficient compute frontier. This is basically the accuracy of the system to correctlt identify targets. This is due to the deviance in the algorithm which noise (other objects), introduce. So, even if an object is not obscured, if there is sufficient noise in the object, a lock may not be possible.
This has no known solution, by the way. The only way to meaningfully improve the accuracy of such a system is just through raw power. More reference data allows for noise to be ignored easier. So even if neither flares nor other objects are obscuring a plane, the simple pressence of background objects does actually degrade the seeker’s ability to maintain a lock due to this noise floor. This degredation of course depends on the processing power of the missile. Which can be scaled fairly well based on age. But regardless.
Also, similarly to this, its not that the whole plane has tk be covered, only a portion of it, again tied to deviance from expected. Ie newer missiles will handle it better.
Well at close ranges, even fairly old systems can reliably detect targets at close ranges with fairly high background noise. It should also be noted that this is very dependent on the target’s resolution on the sensor. The fewer pixels there are for the plane, causes the error rate to drastically go up. As this lowers the noise floor for error.
Further compounding the range issue is that stuff like the shannon limit does actually apply to the target’s optical signature due to atmospheric distortion. This is more of a hard limit as unlike the other range limiting factor, narrower FoVs dont help against this. But this is relatively a minor thing untill you’re looking at very long ranges.
None of these issues alone are enough to defeat IIR systems. But any combination of them? The more these come into play the worse the noise floor, the more likely the missile is to fail to track. Although it should be noted modern flares are still mostly needed for this, as older ones produce too much bloom, which is relatively easy for edge detection algorithms to work out due to their smoothness.
And also depending on the exact type of image detection there are other caveats. Ie some systems dont use pre-trained target sets, instead tracking based off of an initial target lock. Which makes them massively susceptible to pre-flaring, but otherwise has some benefits which wouldnt mean much in WT.
8 Likes
How do you know all this? Just curious btw, no personal attack or anything.
For this specifically, a combination of reading papers on missile guidance, and my own personal experience making image recognition algorithms.
6 Likes
Pretty cool info ngl. Thanks for sharing your knowledge.
it is a hud repeater using a small camera that is facing forward
One fella sent me a dm with explanation, but still thanks for the info.
a dual pulse meteor will probably be worse than aim120 lol .
That’s… Quite impressive.
absolutely, you cant just model variable burn as dual pulse, its going to kill the ramjet, on the other hand if modelled properly its going to be extremely good
Thats how they will model it though, Like an R-77M and an Aim120D
It doesn’t dual pulse.
Boy ain’t it hard to read stuff through fully or something
Making worth reading trough fully and will gladly do so
It’s not my job to appeal to your hilariously small attention span, go watch brainrot or something if you’re incapable of reading more than 3 words
If you are engaging in a conversation with me - it is very much your job to do so. You could’ve ignore my comment completely and yet you didn’t. My attention span is pretty healthy, but only towards things worthy of said attention. Right now you are failing miserably to reach this, actually pretty low bar.
Highly questionable, you seem incapable of actually arguing in good faith or discussing anything if you only care to read things that “interest you” (assumed to be things you agree with, considering previous evidence)
You are free to assume or question anything, doesn’t mean your assumptions are correct. Arguing with good faith over the Internet is a privilege, not a given. I have zero reasons to engage in a good faith argument with you
