Dear Vikram, It is not good to calculate NDVI from DN. The purpose of DN (digital number) is only for packing as much information as possible into a given âBIT-sizeâ (in Landsat 8-bit/pixel/band). DN is practically the radiance (I think TOA?) modified with an offset and a gain in order to press or expand the range between 0 and 255. This means that the relation between bands is changed, and this is a crucial need for NDVI (and generally for indices). Probably the minimum pre-processing you absolutely need to do is the conversion from DN to RADIANCE (TOA) using the gain and bias information in the MTL file (metadata). Better and more precise is to calculate the ground reflectance. I did this with the RemoteSensing package on R-Forge. But the current version doesn't work because of incomplete modifications in the package code. Non-vegetated areas can have positive values. I think the soil has often a NDVI of 0.1. Clouds sometimes have even 0.2 as...anyway you can't trust a NDVI from DN and you can't sharply differentiate veg/nonVeg by positive or negative ndvi values. Water normally has a clear negative NDVI (~-0.2). Also implemented in the RemoteSensing package are functions to compute other vegetation indices, I did not test them but maybe you can find something useful there (type â?ndviâ after installing and loading the package). Cheers Matteo
>>> Vikram Ranga <babuaw...@gmail.com> 1/18/2012 10:49 am >>> Hello list, I am working with landsat images and would like to create a phenological cycle with NDVI. I am using R for that but NDVI values changes if i use on sensor reflectance instead of DN values i.e. giving different result in every case. I tried to compare histograms seems to have changed and I am not getting negative values where there is no vegetation but if i use DN values it gives negative values to such locations. Do anybody have any idea what went wrong?? Thanks in advance and would be much appreciated. Vikram _______________________________________________ R-sig-Geo mailing list R-sig-Geo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo [[alternative HTML version deleted]]
_______________________________________________ R-sig-Geo mailing list R-sig-Geo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo
