| 30 Jan 2016 03:37 AM |
|
|
EVI range is from -1 to 1. When i performed it in ENVI, its giving me the range apprx. -390 to 390 at Landsat 2000 data, with the highest peak between -2 to 2 apprx.
Whats this anomaly for?
|
|
|
|
MariM
Veteran Member
Posts:2396  
| 02 Feb 2016 12:44 PM |
|
|
Were you sure to calibrate your data to reflectance, which will range from 0-1.0? This is the expected input for the Spectral Vegetation indices.
|
|
|
|
Deleted User
New Member
Posts: 
| 02 Feb 2016 09:05 PM |
|
|
Yeah... i worked on it... the value starts from around 0 but it cross the 1 ... Should the values greater than 1 are considered as obsolete or error?
|
|
|
|
MariM
Veteran Member
Posts:2396
| 03 Feb 2016 10:07 AM |
|
|
You could get some over/under corrections. For pixels that exceed the expected range, what do you see for reflectance values in those pixels?
|
|
|
|
Deleted User
New Member
Posts:
| 05 Feb 2016 02:52 AM |
|
|
Mari, for performing EVI, did i still need it to calibrate to Reflection? Why so? Cant i directly used the Spectral Reflectance Bands to calculate the EVI...? Altough i did, and i encounter the same error for having values exceeding the +/- 1 on either sides.
|
|
|
|
MariM
Veteran Member
Posts:2396
| 05 Feb 2016 07:00 AM |
|
|
The expected input to the Spectral Indices is calibrated reflectance that ranges from 0 - 1.0 (approximately) which is what you would get if you used Radiometric Calibration to calibrate to TOA reflectance. If you also used FLAASH/Quac for atmospheric correction, then the output is scaled by 10000 and would need to be converted by dividing the result by 10000. If your data is already reflectance within this range, then that is the expected input to the tool.
|
|
|
|
Deleted User
New Member
Posts:
| 06 Feb 2016 04:07 AM |
|
|
Why i have to calibrate it to TOA as i have to used Surface Reflectance for the calculation of EVI, i guess so... EVI work on Surface rflectance bands rather than on TOA.
|
|
|
|
MariM
Veteran Member
Posts:2396
| 08 Feb 2016 08:28 AM |
|
|
You can use either TOA or surface reflectance as long as the data are in the expected range of 0-1.0 (0-100%). Typically data is scaled and a radiometric calibration is necessary to get to physical radiance units before converting to surface reflectance which is why I mentioned Radiometric calibration.
|
|
|
|
Deleted User
New Member
Posts:
| 08 Feb 2016 08:43 PM |
|
|
But the data is already in spectral refelectance, as directly downloaded by the Earth explorer, that's makes me wonder.
|
|
|
|
MariM
Veteran Member
Posts:2396
| 09 Feb 2016 06:22 AM |
|
|
Then it may not need to be corrected. If you calculate statistics on the data, do they range from 0-1.0?
|
|
|
|
Deleted User
New Member
Posts:
| 09 Feb 2016 09:06 PM |
|
|
Nops... it ranges from -9k to + 9k....
|
|
|
|
MariM
Veteran Member
Posts:2396
| 10 Feb 2016 12:30 PM |
|
|
If this is how the data was distributed by the data provider, then it looks to be scaled reflectance to me. I would ask them how to convert it to % reflectance which typically ranges from 0-1.0.
|
|
|
|
Deleted User
New Member
Posts:
| 10 Feb 2016 08:26 PM |
|
|
Yeah... it worked. Thankyou... scale factor of 0.0001 is used by the USGS, multiplying with this, will solve the issue.
|
|
|
|