If you read the books on pollen identification, you'll find the authors used the same process you describe here. So they'd be the first ones to suggest that you do the same, especially since the plant species they have identified most probably won't include all the species found in your local area. There is also the problem of variation in colour shading in pollen pellets from the same species, depending on time after collection, and addition of nectar. One of the good things about the Kirk book (A Colour Guide to Pollen Loads of the Honey Bee) is that it uses a standard colour chart (recorded as CMYK screen tint percentages). If you really want to get specific about pollen colours (and their variation), using such a colour chart to compare pollen loads collected in your area is a good idea. A chart comparison may also help you differentiate pollen loads from different species with very similar colours.