I am looking at the eval version of NI Vision Assistant and from what I can tell is that you need to use "IMAQ ExtractColorPlanes". This will give you three images to work with. One for each red, green and blue. You can analyize each image seperatly or recombine them to one image with an operator (add for example).
The eval of vision assistant shows an option for the "Extract color planes" that give the "value plane". Since I only have the eval version I cannot create the labview vi to see the code it would create, but I assume that it does an addition of the three images producted by the extract color planes vi to get one image of just grayscale.