What considerations should be taken into account when trying to derive a NDVI from an urban environment?

I pulled a tree canopy layer from LAR-IAC data and used an NDVI of 0.38. However, this FAQ on Vegetation in Remote Sensing recommends using >0.8, while I've read >0.6 as a good NDVI for dense vegetation. Is 0.38 too low, and is it possibly lower because of an urban environment (Pasadena, CA)? Should I be using an alternative index, like SAVI (Soil Adjusted Vegetation Index), but maybe for urban areas?

  • 1
    NDVI should produce a continuum of values ranging from -1 to +1, are you trying to come up with a threshold value for determining the presence of vegetation? What RS instrument is it from and at what scale? – scw May 15 '11 at 5:47
  • The threshold value we came across was 0.38, but it seems low, according to the FAQ I linked above. I was curious if it was due to running a NDVI on an urban area, and maybe there are correction methods? The data was categorized as Color Orthogonal Imagery, 4” resolution (urban areas). This pdf f/ LAR-IAC says Pictometry did the data gathering f/ a plane, but not what instrument. What aspect of NDVI does variance in the instrument affect, given NIR and red wavelengths? – htomita May 15 '11 at 9:31
  • 2
    The reason I ask is because most orthophotos only contain three bands in the visible spectrum, and don't have a NIR sensor when doing the flyover. Also, which wavelength the Red and NIR can help figure out the data's correspondence to other more well established uses of NDVI. – scw May 15 '11 at 19:26
  • ah, good tip scw. So you suggest that if I find out the wavelengths of the Red and NIR from the flyover, I can then compare with other case studies? Yes, I don't know the exact range of wavelengths, but I will find out tomorrow. – htomita May 16 '11 at 4:47
  • 4
    have you found out anything further about the data? – scw May 21 '11 at 6:19

Definitely there is a bias in Urban area NDVI, I would prefer to use enhanced vegetation index (EVI)indices for urban areas. The EVI is an 'optimized' index designed to enhance the vegetation signal with improved sensitivity in high biomass regions and improved vegetation monitoring through a de-coupling of the canopy background signal and a reduction in atmosphere influences.


I'd compare your imagery to known values, like using side-by-side web maps and come up with your own threshold. Depending on your imagery, time-of-day, time-of-year, weather, can all impact your result.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.