Hmm, good point; I should have specified opacity to species (I've edited the post in this sense).
But not necessarily. One could in principle cast a shadow without directly blocking the light... some strange effect like a reverse Fresnel spot. You still block some ambient light, but some light still goes through you?
Otherwise ... what happens if you look at an invisible object that lies between you and the sun (or other bright light source)? It would certainly be detectable to normal sight.
Certainly! That is the consequence of the target still casting a shadow, and a limitation of VoI and PeIm invisibility.
But that runs against what Veil of Invisibility says: "The target becomes completely undetectable to normal sight, regardless of what he does, but still casts a shadow."
That it "still casts a shadow" obviously makes a target detectable, if it is not avoiding casting its shadow onto onlookers.
When dealing with Mythic Europe physics, making a distinction between what affects species and what affects light is absolutely necessary.
Quite. So opacity might just not be the right term. No other problem.
Keep in mind that in Mythic Europe our eyes do not react to light, they react to species.
Also keep in mind that the rules of Mythic Europe do not actually match our reality.
Light affects eyes indeed. ArM5 p.140 CrIg Flash of the Scarlet Flames can even permanently damage eyesight. But mainly light carries iconic species.
Actually, you make it simpler than it is.
Forget the sun. Let's say there's a square room. You stand at the center of one wall, one eye closed, and looking with the other at a candle burning on the wall opposite to you. Two large bonfires are shining bright to each side of you.
You can clearly see the candle. Now, you hold a very small twig at arm's length in front of you, aligned with the candleflame, so that it partially eclipses the candleflame. (I've actually seen kids do this a lot, the twig looks as if it's burning with the candleflame).
A mirror can clearly tell that there's no detectable shadow on your side of the room, either on you or on the twig - remember the bright bonfires? So far, the experiment is mundane - you could carry it out at home.
Now, suppose a magus casts Veil of Invisibility on the twig. Do you visually detect it? According to you, the only way you'd visually detect it would be if it projected a shadow, but there are no visible shadows on your side of the room, so you cannot detect it. On the other hand, you say that it blocks the light of the candle, so you should only be able to see the "outer" part of the candleflame and visually detect the twig silhoutted against it. But that violates what Veil of Invisibility says it does.
Not so, if there's no room to go around it.
Consider one-room hut, completely sealed, with a nice (magical) fire making the inside warm and illuminated. A magus inside the hut casts a variant of Veil of Invisibility on a hut's wall so he can see outside and enjoy the view, without experiencing the cold. If the species from the outside couldn't make it through the invisible wall, the magus would not be able to see anything.
Actually, you cleverly and carefully positioned the twig in a way, that its shadow from the candle fell into your eye. Right? 
Do species need room? Like scholastic angels?
In the sense that they can't enter a space completely surrounded by materials blocking light? Yes, they do need an opening, however narrow (remember, we are tatlking of optic species here). Otherwise, if the can enter an illuminated space sealed to outside light, someone inside ought to be able to see outside.
Wrong! As I said, and as anyone can check if you close your eyelid, there is no visible shadow cast by the candle.
Note, that the iconic species in question are moved by magic and not carried by light. Of course magic would then have to deposit them in light again on the onlooker's side.
Sure. But see, you were claiming that "It is also quite irrelevant ,whether the [optic] species are carried through or around the target." I said that it is relevant, because if there's no room around the target, the only option is for magic to carry species through it. Which indeed matches your words, when you say that, to move them into an enclosed space, magic would have to deposit them "on the onlooker's side".
Let me make it easier. Suppose there's a spherical room, the inside having a radius of 2 paces, and the outside a radius of 3 paces. The 1-pace thick wall is impenetrable to light. In your thought experiment, species start outside of the wall (so at a distance at least 3 paces from the centre of the sphere) and reach said centre. Are species ever at a distance between 2 3/4 and 2 1/4 paces from the centre of the sphere?
If they are, they are going through the wall. If they are not, the magic is teleporting them, which is neither making them go around nor through the wall.
To join in the fun, on ArM5 p.79:
Imaginem spells affect the process by which species are produced, rather than the species themselves.
Thus, anyone saying that "destroyed species can't be detected because they where destroyed, there's nothing to detect" isn't quite grasping it.
The Discerning Eye doesn't care about species, as far as I understand. It cares if there's an illusion caused by a spell (check the guideline).
"Illusion" is an ill-defined term in the book, but nowhere, as far as I managed to check, it says invisibility is not a kind of illusion. It could be construed as it not being (by a reading of Second Sight, the invisibility spells, the InIm spells, etc), but that's not given.
If we classify illusion as "any effect affecting the process by which species are produced", then Perdo (which alters the process to the effect that species are not produced anymore) is an illusion. If we, otherwise, define illusion as "images affected by Creo, Muto or Rego" then, of course, Perdo is automatically excluded. If we define invisibility as "not a type of illusion" than Perdo is also excluded.
There's no particular reason, however, to classify a PeIm effect as invisibility, but a MuIm effect that shrinks the image of the target until it's the size of a grain of sand as not-invisibility. Invisibility is a descriptive thing (what I can't see is, by definition, invisible) not a normative thing (only such and such are trully invisible).
It's also key to dissociate "created or altered through a spell" in the description of Discern the Images of Truth and Falsehood and The Discerning Eye from "created through Creo or altered through Muto or Rego", which is what a few people are reading. Otherwise you have already made your judgement, and your arguments are just seeking to support that.
In previous discussions in this forum I have been on the camp that The Discerning Eye could not detect Perdo invisibility, but after temprobe's first post and detailed consideration I found my previous arguments failling to support my previous position. It's not that The Discerning Eye or DtIoTaF surely see through PeIm effects. It's just that there's no real reason for which they shouldn't be able to, as far as I can see (pun not intended).
A few things have been said about how species must be redirected, specially on how an invisible person casting a shadow is evidence that light carries species, and thus the species behind an invisible target must be carried or displaced to go through the person.
That's a way of thinking that doesn't fully agree with the rules. It would mean that Creo images can't block the view of whatever is behind them (since created images have no substance, and thus, don't block light, nor produce shadow). One could posit that every Creo Imaginem spell holds in itself a seed of Perdo (but w/o requiring requisites), to destroy what is behind? And makes this in a directed way, depending on where the observer is? This is clearly speculative and unsupported, non-sense I'd even say.
Imginem is weird. The fact that Imaginem doesn't agree with our modern understanding about the senses, and the fact that our thought experiments can't provide much insight because we have little basis in the real world with which to compare it make it weird and abtruse.
It's best to not dwell too much on the specific mechanics by which a created image blocks other images behind it, or why a destroyed image allows one to see what is behind it. The metaphysical (or rather, metamagical) reasons on which the phenomenom relies are hardly important, and each person will have a different understanding about them.
My personal one is that CrIm secretely creates an intangible, but not invisible, ghost inside the image to destroy the species, and PeIm links the image to an abtruse, small and empty magical regio, sending the species there (making it actually a weird combination of Rego and Muto, but still using the Art of Perdo) while also making the body transparent to the species behind (with Perdo, of course) but still tangible (needs Muto). It's all very complex, you'd need a Magic Theory score above 16 to fully grasp it).
The bottom line is: the relevant thing for the discussion should be the guidelines (and other rules related to images and detection of them), the descriptions and practical examples on hermetic magic, and when useful, historical understanding of the phenomenom (on which the guidelines generally, but not always, are based). Speculation on the metaphysics of the effect has no place at all unless the first three fail us.
PeIm stops the process by which species are produced. So, there are no species to detect. For practical effects, not producing them on source it's the same as saying that they are being destroyed on source.
All the other arguments that you made are interesting, but I don't understand what more must be grasped there tbh.
That this:
is inconsequential for the spells under examination. They do not rely on species existing.
If the species are destroyed at the source or if never generated to begin with is a metaphysical exploration that is also inconsequential.