View source for evi

{{DavinciWiki_Function|
Generates an Enhanced Vegetation Index map
|
Arguments: 1- or 2-dimensional greyscale images representing near-infrared, red, and blue spectral bands (all three are required), and optionally several coefficients.
 
Return Value: A [[float]] greyscale image
|
[[Syntax]]: evi(nir, red, blue [, l &#61; VAL] [, c1 &#61; VAL] [, c2 &#61; VAL] [, gain &#61; VAL])

The optional '''l''', '''c1''', and '''c2''' parameters refer to coefficients in the original formula, with '''l''' representing "the canopy background adjustment that addresses non-linear, differential NIR and red radiant transfer through a canopy", and c1/c2 being "the coefficients of the aerosol resistance term, which uses the blue band to correct for aerosol influences in the red band". If not specified, they default to '''1.0''', '''6.0''', and '''7.5''' (the values from the source paper) respectively. The '''gain''' parameter is a simple scaling factor and is included only for complete representation of the original formula. It defaults to '''1.0''' if not specified.

This function is part of the larger class of vegetation-index methods used in satellite image-processing. See [[ndi]] or [http://www.beneaththewaves.net/Photography/Calculated_Greyscale_Images.html this discussion of related functions] for more detailed information.

EVI is of the subtype which will return both positive and negative values. Positive values generally indicate the presence of vegetation (with greater values indicating healthier vegetation). Negative values generally indicate a lack of vegetation (water, rock, soil, etc.).

The formula used is:

'''EVI &#61; G * ((NIR - Red) / (NIR + C1 * Red - C2 * Blue + L))'''

This formula was obtained from "[http://www.oneonta.edu/faculty/baumanpr/geosat2/Central%20Africa%20II/Central%20Africa%20Part%20II.htm Tropical Wet Realms Of Central Africa, Part II]" (Paul R. Baumann, 2009).

Note: as the example below indicates, this formula generates results over a ''much'' broader range than most other vegetation-index formulas. As a result, even when the results are normalized, they are "grainy".
|
 dv> nir
 2000x1000x1 array of float, bsq format [8,000,000 bytes]
 
 dv> r
 2000x1000x1 array of float, bsq format [8,000,000 bytes]
 
 dv> b
 2000x1000x1 array of float, bsq format [8,000,000 bytes]
 
 dv> result &#61; evi(nir, r, b)
 Generating an Enhanced Vegetation Index image.
 2000x1000x1 array of float, bsq format [8,000,000 bytes]
 
 dv> max(result)
 1414549.625
 
 dv> min(result)
 -1644825.125
|
Full set of LandSat 7 imagery:

[[Image:NYC-Band_Comparison.PNG|300px]]

Output of evi():

[[Image:Vegetation_Index-NYC-EVI.PNG|300px]]
|
|
*[[arvi]]
*[[asvi]]
*[[avi]]
*[[gemi]]
*[[kt1]]
*[[kt2]]
*[[kt3]]
*[[kt4]]
*[[kt5]]
*[[kt6]]
*[[msavi]]
*[[msavi2]]
*[[ndi]]
*[[slavi]]
|
*dshadow.dvrc
|
1.00
}}
[[category:Functions]]
[[category:Landsat Science]]
[[category:User Defined Functions]]
[[category:Data Manipulation]]
[[category:Davinci's Shadow]]