in an equation for emissivity calculation i should to determine the value which is exactly boundary between soil and vegetation and the value which is placed between mixture pixels(soil+vegetation) and full vegetation ones, on NDVI image using histogram.
Generally, speaking for NDVI, or any other indices, nothing is that accurate. There is always compromise. it depends upon your required level of accuracy.
One can use thresholding techniques to differentiate. Simpler one is OTSU (i.e. Matlab code). Atleast it can give you an idea of some value. you can then check it by going above and below that threshold.
you can see the methodology of this article for distinguished different NDVI values:
https://www.researchgate.net/publication/309589918_Driving_force_for_forest_fragmentation_explored_by_land_use_change_in_Song_watershed_India
Article Driving force for forest fragmentation explored by land use ...
Which equation are you using to calculate emissivity? Do you have any reference that we can have a look at? Are you using emissivity for Land Surface Temperature (LST) estimation?
If you mean the NDVI based emissivity which Sobrino et al. (2004) talks about in his original research paper (link attached), they use NDVI < 0.2 as bare soil, NDVI > 0.5 as fully vegetated and NDVI values between 0.2 and 0.5 as a mixture of both. However, you should keep in mind that this is specific to a particular location and it may be different for your study area. In any case, have a look at the paper.
http://www.uv.es/ucg/articulos/2005/Publications_2004_10.pdf
- Badges
- Science topic
- Similar topics
- Methodology
- Laboratory