The cleanest application is fluorescence imaging where the field of view has a brightness gradient — illumination variation, broad autofluorescence, or any background that varies on a much larger scale than the cells you want to detect. The cells are bright spots; the background is the slowly-varying canvas underneath. Top-hat separates the two cleanly.
Use a structuring element a few cell-diameters wide. The cells (small, bright) survive the subtraction; the gradient (large, slow) is captured in the opening and removed. A single thresholdLoading... on the top-hat result then operates on a flat baseline, and the same threshold works across the whole field.
The same logic applies in brightfield imaging when small dark features (dots, granules) sit on a varying brightfield canvas — there, the symmetric tool, bottom-hat, separates them in the same way.