Which one of the two methods of "Spectral Mixture Analysis (Linear Spectral Unmixing)" and "Using the linear/quadratic model to relate NDVI to Vegetation Fraction" gives a better estimate of vegetation fraction and thus percent vegetation cover? Any suggestions would be much appreciated.
In the Name of God
hi,
1. Due to map/image and study area
2. The purpose of your research
3. review of both methods in studies
Sorry I know little about it
if you have multispectral data you may try to apply unmixing, even if it would be better to have info on spectral characteristics of soil in the area. However, if you have a limited number of spectral bands (i.e., only red and nir) vi-based relationships are the only suitable alternative.
Thanks Carmelo. Indeed, I am working with landsat imagery, thereby, I am having multispectral, however, the soil data is lacking! So in this case, I may better to use linear modelling to relate NDVI to fraction of vegetation and then calculate the percent cover? Would that be sound? How do you see about the accuracy of each of these two methods?
- Badges
- Science topic