In the brewing industry we use Kieselguhr (perlite) filtration with filters like this: https://www.ryebeckltd.com/current-equipment/schenk-kieselgu.... These can sterile filter particles as small as 1 micron and are density-independent. These work by having a steel plate with fine slots upon which you deposit a coarse bed of filter material, and then you dose kieselguhr into the liquid flow passing through, and this gradually builds up on top of the coarse bed to create an ultrafine filter. The ultrafine filter increases in depth over time, and this prevents blinding (blocking) of the filter as filtered particulates accumulate. While used to remove yeast and other particulates in brewing to create "bright" beer, it can be applied to most liquid filtration problems, and at high flow rates (30000 litres/hr); and you can run multiple filters in parallel to further increase the flow rates. We also used kieselguhr in research labs for fine filtration--just add the powder to regular filter paper.
In practice this is based upon manual calculation and adjustment. In the place I worked, the filter room had two of these, each about 1.5 storeys high. It had a dedicated control room to operate, and the two operators would be continuously monitoring and adjusting the flow rates, dosing and other parameters to ensure efficient operation. When you have an unexpectedly dirty product or you muck up the calculations, the flow rates can become very low and that starts to cost serious amounts of money when it's a rate-limiting step in an industrial process.
When the filter becomes too inefficient, the backpressure excessively high, or the filter depth approaches the tank capacity, the metal filter plate is spun and this destroys the filter bed which is spun off. The coarse filter base is then reapplied to "regenerate" the filter.
You can filter them - and they certainly do - but with sea and lake salts the filtration is mainly to screen out sediments such as sand grains and bits of shell. Naturally, the filtering is a costly process and the filters need to be discarded after a certain level - I imagine there is a cost-control factor here, where screening at a 1mm threshold catches 99.9% of sediments to make the product reach a certain purity level and more thresholds are not worth the money.
Further screening is probably not cost effective - and seeing from the results of the study, plastic microparticles are not a source of concern.