Giving a little bump to this topic because I wanted to point out some of the many parameters that go into to getting the best sensitivity for minor planet detection. If you want to actually have a hope of discovering an asteroid these days with modest amateur equipment (14" aperture or less), then you need to maximize target photons. That means, no Bayer masks (i.e. color cameras). On a pixel by pixel basis, a Bayer mask kills more than 65% of your signal. That's like swapping an 11" aperture for 6.5" (35% of the photon collection area). You must use a mono camera.
You also want your mono camera (these days, almost everyone is using CMOS rather than CCD) to be back-illuminated and cooled in order to maximize quantum efficiency while maintaining low read noise and low dark current. Back-illuminated CMOS devices can have a quantum efficiency peak over 90%, compared to front-illuminated that rarely tops 60%.
Another factor is image scale: you want to marry the right focal plane to your optics (specifically the focal length) and your typical seeing. Critical sampling is about 2-3 samples of your FWHM in order to get good astrometry (positional accuracy) while not smearing your signal over too many pixels and thus losing sensitivity.
For instruments with apertures greater than about 5 inches (the only ones with any hope of discovering new asteroids), your asteroid and star FWHM sizes will be driven by your astronomical seeing and the quality of your tracking. With a good mount and good guiding, you're ultimately limited by your seeing. For example, if the best seeing you typically experience is 1.5", then you'd want to target an image scale of 0.5" to 0.75". Since you've always got the option to bin 2x2 or even 3x3, it's perhaps better to be optimistic about your seeing and lean toward the smaller image scale.
Here's an example. I've got a Celestron EdgeHD 11" telescope, and for reasons that will become apparent, I'm using a 0.7x focal reducer which drops the focal length from 2800mm down to about 1960mm. I went with ZWO's ASI 294MM Pro (mono camera) because of its high QE, low noise, good well depth, and most importantly a pixel pitch I could work with. Initially, camera users did not have access to its native resolution (8288 x 5644) but only its internally binned 2x2 data (4144 x 2822). The effective pixel pitch was then 4.63 microns rather than native 2.315 microns. So my 2x2 binned pixel scale is 4.63 microns / 1960mm = 2.362 microradians, or 0.487". That's perfect for 1-1.5" seeing. (Notice that without the focal reducer, my image scale would be 0.341", i.e. 3x oversampled for 1" seeing. I'd rather have a 43% larger FOV than be that oversampled even under great seeing.) On the extremely rare days where my seeing improves to 0.5", I can go to that camera's native resolution.
The final puzzle piece, and the main reason for my post, is exploiting the spectral content of my target, while (if possible) suppressing the wavelengths most impacted by light pollution. Asteroids tend to be "red" in that they have higher solar reflectivity from yellow to near-IR than they do in UV to green. Since there are significant light pollution sources at some of the shorter wavelengths (e.g. high- and low-pressure sodium vapor), it certainly makes sense to filter these out if possible. Unfortunately, LED streetlights are now also a big source of light pollution, and while they peak in the blue, they also have significant spectral content out to 650nm (red), but rapidly diminishing beyond 650nm.
It would seem that from light-polluted imaging sites, what would be optimum is a long-pass filter that cuts out everything below, say, 630nm, but passes everything else out to at least 1000 nm (1 micron). Astronomik's CLS filter is close to doing that: it has great transmission beyond 645 nm and notches out everything from about 540nm to 630nm, which covers the worst of the sodium vapor light pollution. But it has a second transmission window from 450-540nm which lets through the peak of the LED street light spectrum, where it might be better (from a signal-to-noise ratio perspective) to chop out *everything* below 640nm if asteroids are what you're after. I get the motivation for having a transmission notch there in the CLS filter: it lets through the blue-green hydrogen beta line and the green Oxygen III lines, which you certainly care about if you're doing DSO imaging. But if I'm chasing minor planets, I don't care about those.
So ... is there a filter out there that cuts everything below 630nm or 640nm, but is otherwise transparent out to 1 micron?
