I now have found found multiple mainstream sources that agree with sol in saying the average density of the universe is one hydrogen atom per cubic meter (whether that includes dark matter and dark energy as sol seemed to imply is not clear). But let's go with your 1.6 per cubic meter number. It makes no difference to the problem I see.
Average mass/energy density from sol, average
mass and
baryonic mass from me. But I might have gone a bit wrong with calculations, it's not unknown, and I'm happy to be corrected by anyone that knows better (heck, once I calculated the mass of the observable universe as 30kg.... think I lost about 8 factors of the speed of light though!)
But not on the order of a 1000 to 1. That's not what I see in images like these which come from various mainstream simulations and real data:
[snip]
http://astronomy.sussex.ac.uk/~mif20/images/millenium.jpg
[/snip]
OK, I was thinking of the Millennium simulation at the time. I strongly recommend you look at a Millennium sim video because even if you don't believe it is at all like our universe it a) gives a better idea of the 3D structure and b) is very pretty. You've given an image very much focussed on a cluster of dark matter there too, and the baryonic matter is biased to dark matter, and that's obviously not a terribly representative region.
I'm not going to comment in great detail of the others - some of the others don't suffer from the peculiarity of focussing on a cluster (I see after writing the above that you've got a much wider view of the Mill. sim there too, so I was a bit unfair but it's still worth bearing in mind for some of the other pics) but you also again have to remember they're a projection through a large 2D volume and this may inflate the apparent density. Again, I'd encourage you to grab a flythrough vid.
Most of the sims do show dark matter densities, not baryon densities, as the bias factor isn't quite as well understood as we'd like. And the scalings in the imaging are not necessarily linear. A faint region may be much less dense than you might think, I'm not sure, but I do expect they're made more to look pretty than anything else - if one wanted to check the results you'd probably get hold of actual numbers or a portion of the simulation in question.
Galaxy survey images will show a sizeable blob for each galaxy, again to make it easier to see. It doesn't necessarily correlate to actual physical densities.
*edit: I should better explain that by bias I mean that the ratio of dark matter to baryons is not the same everywhere. The relationship between the densities of the two is not simple, but there's a general principle that where you have a lot of dark matter you have a lot of baryons, as you'd expect from the gravitational link between them. It doesn't mean that the distributions of the two components of matter are identical.