I am trying to become more familiar with processing NDVI values in GIS. However, despite its usefulness, it seems to have some limitations. Namely, creating NDVI values across a large scale can be problematic if shading (i.e. cloud cover) and other lighting conditions are not constant especially when comparing NDVI values of the same area over time (if I am understanding this correctly). I have local digital images of 1 meter square plots taken from a high resolution hand-held camera (and with a hand-held NDVI camera) of vegetation plots at a height of ~5 feet. I would like to generate an average NDVI value for these images and use them as covariates in a survivial model of my target species. Ideally, I would upload these photos as TIFFs into ArcMap for processing. I have seen people use the red and infra-red bands from Landsat images to generate an NDVI value for the raster layer. Could I apply this to my photos?
Also, is it acceptable to use the NDVI value at each location as an independent predictor variable? Or, will the local NDVI values be confounded by variation in lighting conditions even if I'm not comparing the values from each image directly? I standardized the area photographed, time of day (10am-2pm) photos were taken, and camera settings.