Hello, I need some help for post classification processes. I used Naip imagery to obtain thematic maps of a watershed in Texas. However, to comply with a systematic classification scheme like Anderson I need to do some post classification processes. To elucidate, my classification image has 7 classes having trees, grass, and exposed in particular. However, I would like to aggregate these individual classes in order to get coarser classes like a deciduous forest, evergreen forest and rangeland and so on. In NLCD a deciduous forest is defined as having at least 20% cover of deciduous trees while 75 percent of the trees shed their leaves. The difficulty here is the fact that while I can map individual deciduous trees by Naip imagery since it is a 1 meter-spatial resolution image, I don't know how to quantify and aggregate these individual trees in a larger pixel. For instance, I assume I pick a 30-meter pixel size in order to correspond with NLCD classification scheme, so how can I quantify the number of 1-meter pixels from different classes in a 30-meter mapping unit? Then if I can quantify is there a way to reclassify or appoint new pixels into new classes depending on the quantities of different classes?
