There is a property of optical beams called "beam quality". It is the product of the size of the beam at a beam waist, and the angular spread of the beam as it propagates from the beam waist into the far field. "beam quality" is important because it is a conserved quantity. (there are much more complicated and formal definitions, but that's the basic idea) You can make the beam waist larger, say by putting the beam through a telescope. However, the angles will necessarily get smaller and the product of angle and size will not change. Without throwing away light you cannot make the product of beam size times angle smaller. You CAN make it larger with diffusors or aberrated optics, but let's ignore that for the moment. With good optics and no loss of light, the beam quality is a constant.

Beam quality determines how bright you can make the beam in the far field. Usually you have some constraint on the size of the beam in the near field, for example say you are only willing to make the output of your flashlight 2 inches across or some similar constraint. The beam quality divided by the aperture size is the smallest divergence angle you can achieve. The beam divergence determines the beam size at a distance and the power spread over that size is the radiance. Lasers will often use the term "brightness" (without a specific quantitative definition) to refer to a combination of power and beam quality which ultimately determines how bright (high radiance) the beam can be in the far field through a given aperture.

Edge emitting diodes are generally "brighter" than VCSELs: expanded to a given aperture size in the near field they will produce higher radiance in the far field. VCSELs have greatly improved in brightness in recent years, so they are quite bright. (particularly multi-junction emitters) However, so have edge emitting diodes, so comparing the state of the art in both varieties, edge emitters remain considerably brighter both because they have higher beam quality (at least in one dimension) and higher power (because the arrangement more readily sheds heat)

However, as you noted, VCSELs have an advantage in that the brightness is the same in all dimensions. Edge emitters have different brightness in the 2 dimensions. The emitted beam is only a micron or two with a FWHM divergence of perhaps 30-40 degrees making the beam quality product in that dimension only ~5 mm-mrad (that is essentially diffraction limited. Very good beam quality) However in the other dimension the cavity might be 100-150 um long with a divergence of 7-10 deg and so a beam quality of 12 - 30 mm-mrad, so not nearly as good in that dimension. The first thing to note is that while the beam size differs by a factor of ~100, the beam quality only differs by a factor of 3-5. It requires anamorphic optics (different focal lengths in the two dimensions, usually accomplished by a pair of different focal length cylindrical lenses) BUT with proper optics the two dimensions don't have to be as terribly different as it might first appear. Now, you have two choices: either the beam is elliptical in the near field, not fully utilizing typical round lenses, or it is elliptical in the far field illuminating a 3:1 or 5:1 aspect ratio, or some combination of the two.

In many applications this is not a disadvantage. Because the ground is flat often in the real world all the interesting stuff we want to see is more concentrated in the vertical dimension and more spread out in the horizontal dimension. Alternatively, in the near field, is often convenient to package a lens that has its top and bottom cut off and so is shorter in one dimension. We typically do a little of both: make the aperture something like 2X larger in one dimension, and also make the field of regard 2X larger in the same dimension.

On the other hand, with a VCSEL you can use round optics efficiently and produce a round spot in the far field. HOWEVER, generally it is always possible to find an edge emitter that will result in much higher radiance ("brighter") in the far field than the competing VCSEL.

VCSELs have other advantages and disadvantages. A critically important advantage is that their wavelength doesn't change too much with changing ambient temperature. (that's the advantage that trumps edge emitters and results in us choosing VCSELs for our designs despite the lower brightness) VCSELs can be unpolarized (although in practice mechanical stress results in them not only being polarized, but being polarized in unpredictable and varying ways) Edge emitters are essentially always highly linearly polarized. Whether you prefer unpolarized to polarized would depend on your application. However, I don't think uncontrolled polarization could ever be thought of as an advantage, so if you care you would have to be sure to use VCSELs that really are unpolarized.

P.S.: While the car headlights may or may not be on, I think what you are seeing there is not light from the head lights, but reflected laser light. Through the same mechanism that makes cat's eyes appear bright when you shine a flash light at them in the dark, or the eyes of a deer appear bright when caught in your car's headlights, optical systems will preferentially reflect the laser light back towards the source making them appear brighter than the surroundings.

The sufficiently bright type.

Actually, there are very few long range laser illumination products.

In several projects, I needed a long range illuminator or two. For example, in the space based-laser, I needed a uniform target illuminator that would work at a few thousand kilometers. I also needed a wavefront beacon illuminator to work at a similar range. The more powerful lasers at that time that could be used for illumination were slab YAG lasers. Today, almost certainly I would choose a YAG fiber laser. But use the 1.06 micron beam, or double the frequency and get a 0.53 micron green beam?

The wavelength question is not an easy one. There are losses in converting one wavelength to another with a non-linear optical process, and the quantum efficiency and noise of visible light detectors is better. In the end, the advantages and disadvantages are almost a wash.

For long range illumination, you might be interested in how much the beam needs to be expanded to get the beam divergence angle you need. Since all laser beams diverge due to diffraction, the diameter of the beam, and the wavelength, have a direct effect on how far the beam will work. To get to a few thousand kilometers required an initial beam diameter roughly a half meter or so, a power of a few kilowatts, and the shortest wavelength possible. But, beam quality is another factor. A short wavelength is no good if there are wavefront errors in the beam.

So here is the bottom line: there is no best product for all applications. You come in with a description of the target, how far away it is, and what kind of telescope you want to use to project your laser beam and what kind of detector you are going to use to see the target.

So, I would say the best one I know of is the one my team designed for SBL-IFX. It was never built because 9/11 happened and we used the money to start the Department of Homeland Defense. The best one actually built was for ABL, shown here. The green laser beams are an illuminator laser array.

There were several great comments above.

Also, since time-resolving image sensors are advancing a lot now one may think of use of pulsed illumination and time gated imaging.