Have you already tried using a microplate reader? Theoretically it should generate some results, but I don’t know if the errors can be assessed. Could you try perhaps an image analysis software? And use as inputs images?
Someday, perhaps, but not in my lifetime, I'm afraid. I do plan to use a "particle" counter to verify the concentration of added exotic pollen (for calculating absolute number per CC), but I'm not aware of any reliable method to identify individual grains with certainty. The current method of suspending a portion of a sample in silicone oil (about 20,000-50,000 centistokes) and counting a fixed number of fossil pollen grains (along with the number of exotics (I like to use Eucalyptus) by traversing a cover slip a number of time until the fixed number of total pollen is reached (excluding exotics). The silicone oil is of such a viscosity that it does not "bleed" all that much outside of the cover slip, and it allows on to lightly touch the cover slip with a probe to move the grains into a different position for easier identification.
In the mid-60s, my mentor at Univ. Mich. (Wm. Benninghoff) was working with some researchers to use holography to more quickly identify grains in the counting process Ciprian's suggestion of using image analysis software may be the route taken, although his point of input images could be problematic in that we "pollen-peekers" often need several views of a particular grain for proper identification. This can easily be accomplished by tapping the cover slip and allowing the grain to rotate in the viscous silicone oil.
I like to stabilize the cover slip with a drop of clear fingernail polish at two corners in order to keep the cover slip in place. The other advantage of using silicone oil (vs glycerin jelly, for example) is that the oil does not dry out over time.
Good luck in seeking a current method and be sure to let the rest of the "pollen-peeking" community know of your findings.
Recently automated method of measurement of ommatidia was developed using ImageJ software (link below). I believe it is possible to similarly automate pollen grains counting, however a microscope still will be needed. The authors of "ommatidia method" say that this method is 70 times faster than traditional manual way while being equivalently accurate. Why not ask the authors about the pollen?
Article Automated Measurement of Ommatidia in the Compound Eyes of Beetles