This isn't going to work that well (other than as a first pass) as many flowers are very similar, and a large number of plants require you to look at more than the flower to separate them.
Portable DNA sequencing is the future, and it will probably reveal all sorts of interesting data about plant sub-species and population relationships.
You can already mow a meadow, whizz up the cuttings, and identify all the plants from the juice. Great Crested Newt surveys can already be done by taking a water sample and sending it off to check for environmental newt DNA instead of (much more expensive) trapping and night-time searches.
It surely won't be that long before your smartphone (or a similar sized machine) will able to do sequencing in the field.