First there's a striking lesson in star temperatures and their effect on the interstellar medium. Have a look at the circular glows around the stars in upper Scorpius: gold, blue, and "raw meat" (I'll call this pink), and maybe two glows with blue/pink mixed. We all know that Antares is a cool supergiant star, with essentially no ultraviolet photons to fluoresce the gas in the rho Ophiuchi cloud (the bright nebulae in the head of Sco, and the three-forked dark clouds extending to the northeast). So all Antares does is light up the dust with its dominant yellowish color. The stars flanking Antares, sigma and tau Sco, are early-B type stars, hot enough to fluoresce the gas, and (bingo) their clouds glow raw-meat-hydrogen-alpha pink.
Just above Antares is the blue cloud around rho Oph itself. This is a pair of stars of type B2, just barely not hot enough to fluoresce the gas, so we see a reflection nebula.Likewise up above, stretching on either side of nu Sco (type B2IV), another blue reflection nebula. beta Sco appears to have no nebulosity associated with it, but both delta Sco (B0.2IV) and pi Sco (B1V) are hot enough to give at least some emission. Their clouds look somewhat purplish on my screen, so perhaps the reflection component is strong enough. Just from the colors of the nebulae and the spectral types of the involved stars, we now know where the cutoff is for an emission object---you gotta have stars earlier than B2. The other lesson that's obvious is that dark/reflection/emission nebulae are not different types of objects in the astrophysical sense: the make-up of the clouds, their composition, is identical, and the only differences are the temperatures of the stars (if any!) in them. No stars and dense = dark nebula; stars of type B2 and later = reflection nebula; OB stars = emission nebula. All three are exemplfied in the same cloud near rho Ophiuchi; here's the sky telling us how the World works!
Now the rho Ophiuchi cloud is only a couple hundred parsecs away, and any interstellar extinction you see in the area is caused by it, not something in the foreground. Thus the pink color of the emission objects represents "unreddened nebulae". Look around elsewhere in the photo at the various nebulae: the big region south of the Norma starcloud (at the bottom of the picture) is a lot redder---but could this be because that area was reddened by Earth's atmosphere? (As the joke goes, most of the "interstellar medium" is within the first mile of the telescope.) The zeta Sco nebula is a healthy pink (unreddened), but our friends NGC 6337 and 6357 are also dark red, and we know that these objects are indeed obscured behind those dark clouds that surround them: reddened red nebulae. But just next door, that nebula just west of M6---which I never knew existed---is bright pink again, and unreddened. The main cluster there, NGC 6383, contains several O and early-B stars, which are doubtless responsible for the UV photons.
One can play the same game with star colors, in this case the grand mean color of the starclouds. It looks as though Tony and Daphne have rigged the color balance so that roughly solar-color stars are white, which if there are enough stars to make the image dense, you see as a neutral starcloud color. The region near 45 Oph, just left of center in the photo, shows a nearly unobscured part of the galactic bulge; similar unreddened regions appear near M7, and from there off toward the east edge of the photo. (Indeed, a nearby cluster of galaxies is visible near 45 Oph, since we can see (barely) right out past the center of the Milky Way here.) But now look at the color gradients of the starclouds in the obscured regions, notably along the 'back' and under the 'rump' of Richard Berry's "galactic dark horse" in the upper left corner. Wow! There aren't any blue stars there unless they're in the foreground, limiting it to a few of the naked-eye stars. This is right in the direction of the nucleus of the Milky Way, which has been measured to have extinction of at least 30 and possibly 50 magnitudes (very patchy in any case), so sure enough you don't see it---even a supernova would be completely blocked.
The large network of dark material in the bottom of the photo seems to form a ragged circular outline encompassing the zeta Sco region and the Norma starcloud: not by chance? (I don't know.) I recall Bart Bok writing somewhere that the filamentary nature of some of the nebulae here were suggestive of a supernova remnant, but I think the characteristic emission lines are not present.
Another thing worth pondering is the differences in spatial scale in this image---always a tough problem in conveying astronomy to the uninitiated. The upper Scorpius stars and circular nebulae are very much in the foreground. Normally I (at least) think of a star's region of influence being always quite small on the sky, no bigger than the star images in this small-scale photo. But here we see that it's actually a few degrees across for these guys that are even 250 parsecs away. But down around zeta Sco, something like seven times more distant, the affected area is tiny by comparison (a factor of 50 in area if the relative distances are as I said). The rho Ophiuchi clouds are at around 250 parsecs, but the unobscured starclouds of the galactic bulge are something like 5 or 8 kiloparcsecs away (just considering stuff closer than the galactic center). Antares, a couple hundred parsecs; M4 sitting right next to it, two kiloparsecs!