I’m familiar with coulter counters, which work best for larger cells (mammalian/yeast) and commercially available ones are really expensive and often slide based. That being said, the principle shows up in microfluidics sometimes. Would only work on efflux/subsample on population.
Acridine orange staining is also used for low abundance environmental bacteria, but again, expensive fluorescence microscope. Could work on bulk population, but I’d worry about reagent costs.
I’d say anyone aiming to work in that range is likely fine working in chemostat mode, since population sizes are so low that the media will be largely unaffected by swings in population size.
There’s always concentration via filtering…