A common equation to express the best resolution of an optical system is based on the overlap of two Airy disks
$$ sin(\theta)\approx \theta \approx 1.22 \frac{\lambda}{D} $$
You can ask what size aperture on earth would have a diffraction limit of its own size in LEO;
$$\frac{D}{L} \approx 1.22 \frac{\lambda}{D} $$
$$D^2 \approx 1.22 \lambda L$$
$$D \approx 1.1 \sqrt{ \lambda L}$$
Choosing rough, ballpark numbers; $L_{LEO}=500 \text{km}, \lambda=600 \text{nm}$, I get 60cm.
So the diffraction limit in LEO is about 0.6m for a 0.6m aperture. The relationship is inverse, so it's 0.3m for a 1.2m aperture for example.
That applies to an imaging system as well as to a collimator for laser illumination if you wanted it.
But that's just diffraction. Tracking smoothly and accurately would be quite a challenge. A normal astronomical telescope mount and drive might not have the right accuracy and smoothness while slewing at almost 1 degree per second.
You could build a special satellite tracking drive and I'm sure that's done, or you could use a pulsed laser - basically flash photography. The standard DPSS NdYAG laser at 1064nm could be used directly, or frequency doubled to green 532nm. They come in a variety of powers and typically pulse at 1 to 10 kHz and have pulse widths (in time) of the order of 20ns. You can clock your CCD accordingly and overlay the frames later.
If you use a pulsed laser, it allows you to use very short exposure times centered at the moment the pulse returns to earth, and an optical filter centered on the laser wavelength to reduce background scattered light (including the daytime sky).
Because the distances are so much larger than the coherence length of those - and probably any laser you can buy off the shelf, you can think of the light as incoherent and you won't see any laser speckle.
It's not going to be as bright as sunlight, which is roughly a kilowatt per square meter, but because of the strobing you can integrate for many seconds (effectively).
So after tackling the diffraction limit and motion smearing, you still have atmospheric turbulence. You can't use (at least conventional) adaptive optics, because it is based on looking through the same column of air for at least milliseconds in order for the mirrors to respond, and slewing at this rate, it's not going to work (easily).
However there is the fallback technique of lucky imaging, and since the satellite is an artificial, engineered, functional object an not an unknown heavenly body, you can process, simulate, and deconvolve the heck out of the data to figure out what the satellite would look like without atmospheric effects.
But remember - if you use the sun, you have a half-degree-collimated beam at oblique incidence, and if you use a laser, the beam is far narrower. Mirror-like surfaces (flat shiny sections of metal, solar panels) may appear quite dark if illuminated at the wrong angle, but suddenly appear to flare if the right geometry happens. If it's diffuse, or crinkled, it may be much brighter than shiny surfaces away from specular reflection. This can cause the images to be very different than what the satellite would look like if you were standing next to it.
Caution: Please adhere to all applicable safety measures and legal restrictions. If you are shooing highly collimated lasers outdoors, even into the sky, you can get in a heap of trouble and could potentially do serious damage to people, aircraft, and spacecraft. Don't try this at home!!
