Jump to content

Photo

NDVI of Landsat ETM


  • Please log in to reply
10 replies to this topic

#1 HTC

HTC

    Sergeant

  • Inactive Member
  • 214 posts

Posted 30 January 2013 - 02:31 AM

Hello every body!
When i calculate NDVI index of Landsat ETM image. I use function:
NDVI = (IR-RED)/(IR+RED).
The result is not exactly. The NDVI of Rich forest = 0.2-0.3. Infact NDVI of Rich forest >0.5. I use Erdas, Envi... for calculation but the result is bad. The data of mine: Landsat ETM 2003.
If someone know solution, please help me. Thank you very much.
Regard

#2 rahmansunbeam

rahmansunbeam

    Sergeant

  • GIS Analyst
  • 489 posts
  • LocationDhaka

Posted 30 January 2013 - 03:23 AM

Infact NDVI of Rich forest >0.5.


Where did you find that piece of info?

#3 HTC

HTC

    Sergeant

  • Inactive Member
  • 214 posts

Posted 30 January 2013 - 06:01 PM

These literature research about forest in my country. When i calculate NDVI index in my Landsat image, the result is bad. Value index is low: 0.2-0.3. The rich tropical forest have high value of NDVI (usually: >0.5).

#4 rahmansunbeam

rahmansunbeam

    Sergeant

  • GIS Analyst
  • 489 posts
  • LocationDhaka

Posted 30 January 2013 - 11:54 PM

Ok, it's really hard to say anything from this end coz you are the one doing all the literature survey. There could be hundreds of reason why the result do not match

-the source you are talking about has different methodology with different data in different software,
-the data you are using is over or under prepared
-the area you are calling 'rich forest' has changed during this time
-and many more

If I were you, i'd go visit those places called 'rich' and see if the there is really something. That's actually ground truthing.
  • HTC likes this

#5 Arhanghelul

Arhanghelul

    Sergeant

  • Members
  • 210 posts
  • LocationKingdom of Heaven (RO, east Europe)

Posted 01 February 2013 - 07:28 AM

Have you made pre-processing of the LANDSAT images before calculating NDVI ? ( atmosferic corrections ? ) :huh:
  • HTC likes this

#6 HTC

HTC

    Sergeant

  • Inactive Member
  • 214 posts

Posted 01 February 2013 - 07:44 AM

My friend!
You said true. I have fixed this problem. The result is good!

#7 notlikely8

notlikely8

    Cadet

  • Inactive Member
  • 4 posts

Posted 31 March 2013 - 07:13 AM

Yeah, you should definitely be careful to correctly pre-process all your imagery to at least at-sensor radiance values for ratio-based index calculation...

#8 ricohizon99

ricohizon99

    Private

  • Members
  • 39 posts

Posted 27 August 2013 - 09:43 PM

how to make NDVI classification to polygons? *newbie here. 



#9 peralta2794

peralta2794

    Private

  • Members
  • 19 posts
  • LocationPeru

Posted 29 August 2013 - 06:27 AM

You can use this tool Density slice (ENVI)



#10 3dbu

3dbu

    Captain

  • Members
  • 1,359 posts
  • Locationfrom hell

Posted 10 September 2013 - 04:39 AM

Nice. Thks

#11 dr_s1ngh

dr_s1ngh

    Private

  • Members
  • 29 posts
  • LocationDelhi

Posted 25 December 2013 - 11:03 AM

Dear HTC!! You should first do atmospheric correction of your satellite image using Envi FLAASH or ERDAS ATCOR module.TOA radiance or reflectance would also serve ur purpose. But for getting accurate values of NDVI, you should use reflectance image calibrated by FLAASH module. if U r using ENVI for NDVI calculation using reflectance image.. u should just write the NDVI formula in Band math as:
(((float(b1)-float(b2))/(float(b1)+float(b2))))
where, b1 denotes NIR and b2 represents Red band.
  • Arhanghelul likes this




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users