ground truth for NDVI maps


I would like to ask the Group about so-called calibration of NDVI maps or connection between real numbers of plant growth and received data.

Do you go to the field in order to measure leaf surface area or make any other research (chlorophyll in a leaf) in order to get a ground truth for data, obtained with UAVs?


You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –


    • Hi Cala were you able to accurately measure non productive areas

      ( i am supossing that these would be the areas in red?)

    • Yes, non productive areas are red and you can measure them, sunflower have high relationship with IAF and yield.

  • NVDI is very subjective. Pix4D gives you the ability to tweak the R,G,B and NIR spectra to optimise for your crop but you have to understand just what you are looking for to do this.

    NVDI runs the risk of being devalued by people not understanding this and selling a one sizes fits all solution.

    For myself I am looking fro Agronomists to help facilitate the whole survey, computation and understanding cycle. Just supplying dumb "data" without optimisation for the specific crop is about as much good as a pretty picture of a green paddock.


  • They are both thermal images of a same field, 200 miles west from Buenos Aires Argentina.

    The first one was taken in Nov 2015, the strip effect that you can see, which is not easy to see or quantify from the ground, it is due to an uneven distribution of broadcast granular N fertilizer applied 4 months before .

    The second image was taken in January 2016, it is double crop corn (corn planted after wheat the same year) and shows the effect of soil water saturation on corn.

    To answer your question, I did not perform any analysis. I just walked the field with the image on my phone, trying to correlate and understand what you can see from above how it looks like on the field.

    • Are your NDVI images derived from images captured by UAV?

  • Interesting topic. I would add the questions: why would you want to assess plant growth? what do you plan to do (different) after you have this information?

    See attached a couple of ground truth pictures on wheat and corn:



    • Hi Jose, How do you take the images and analyze them ? is via drone / sattelite (Sentinel-2) , or how ? Which analysis can you perform. Thx.

  • Hi Andrii -

    This depends on what your hypotheses are and what you expect to measure with ndvi. If you are linking biomass you may want to harvest small plots - if you are doing LAI you will want to measure it directly in plots - or for me I was measuring fungal disease severity as average percent leaf area infected within observed plots. Keep in mind - if you plan to do accuracy assessment statistics like kappa - you will need a certain amount of samples for each class if you wish for reliable results. Hope it helps a little!
  • Yes, you need to check at field what you see in photos.

This reply was deleted.