# reflectance map shows reflectance values between 0 and 1

Hi Forum,

The pixels of my reflectance map, built from an RGB camera DJI X5, shows values in red, green and blue. I need that the pixels show reflectance values between 0 and 1.

Is there any way to do this without having to use a radiometric calibration target?

Thanks

Hey Walter,

If you have an RGB orthomosaic, you could separate each band then normalize them using QGIS or Arcmap Raster Calculator. Or you could use Index Calculator to do the same thing in Pix4D. However the refelctance values that you are getting from an RGB camera may not be very accurate.

1 Like

Thank you very much Selim,

We’re making survey comparing imagery taking from smartphones and drone.

I know how to normalize in Arcmap but i wish to learn how do it in Index Calculator…
How can I do it in Index Calculator ? Is there any example, tutorial or tip?

Hey Walter,

That sounds pretty interesting comparing a smartphone imagery and a drone imagery. In Index Calculator you can define your own formula, so you would do something like, Red/255, Green/255 and Blue/255, which I am assuming your images are 255 bit.

1 Like

Hi Walter,

Please remember that not all sensors are suitable for the generation of a useful reflectance map from which other relevant outputs are derived. Hence if your camera is not radiometrically correct what you are extracting is not reflectance since you do not correct the camera characteristics.

If you are using an RGB sensor to retrieve reflectance, please make sure you are doing the corrections listed in the article below before processing:

https://support.pix4d.com/hc/en-us/articles/204894705

Cheers,

Ina

1 Like

Hi Selim and Ina,

In the Index Calculator tab, the reflectance map shows the following values in red band:

660 (nm) ; 0.24 (min) ; 125.50 (avg) ; 255.00 (max)

I built a new formula named red_normalized (red / 255) so that the values were between 0 and 1

However, in the red_normalized.tif they vary from 20.79 (min) to 235.93 (max).

How can I solve this problem ?

Could you post a screenshot here so I could see the statistics?

As a reminder, the values that you are obtaining do not represent reflectance but an arbitrary digital number and is not recommended to use them for your analytics.

Hey Walter,

It is possible to get a wide range like that, I mean it could be a single somewhere. What is the range for Red band when you apply normalization?

Hi Ina and Selim,

I created a new project and re-created all the processing. In this way, I was able to create the normalized RGB bands using the formulas red / 255, green / 255 and blue / 255 and the values were between 0 and 1. Thank you.
I’m looking for a radiometric target to calibrate these DJI X5 RGB bands properly.

Regarding the NIR band, I have used another camera (DJI X3 RGB converted) and a specific radiometric target to calibrate the Red and NIR bands. However, the NDVI has returned low values (-0.11 to 0.16) where there are a lot of healthy plants and soil.Do you have any suggestions to help me solve this problem?

Hi Walter

If your camera is designed to keep radiometric fidelity and you are applying camera only corrections plus a radiometric target the values should be between 0 and 1. Otherwise, this range is not guaranteed.

Cheers,

Ina

1 Like

Perfect Ina,

Thank you very much!

Hey Walter,

NDVI values could range between -1 and 1, all the other bands Green, Red, RedEdge and NIR should be between 0 and 1.

2 Likes

@ Walter Selim is right, my bad I was referring to reflectance values

1 Like

Ina and Selim,

With your support, I was able to understand the problem with the normalization of reflectance maps and vegetation indexes.

However, although the indices constructed with equal area algorithms present a visually adequate classification, I understand that classification is not adequate to calculate the biomass (for example), because the percentages of the healthy and stress areas will always be the same, am I correct? Can you help me ? What algorithm should I use to calculate biomass by NDVI?

@walter

The NDVI analysis might depend on a variety of factors, including the type of tree/vegetation. We specialize in the production of accurate reflectance maps and index maps. The precise interpretation of these index maps might require some outside assistance from an expert.