Manual calibration of sequoia images


#1

We recently bought an EBee drone which can be used in combination with a Parrot Sequoia camera. Among other purposes, we use this drone to take areal pictures of an experimental field site with drought treatment. When our drone unfortunately crashed, we used the Sequoia camera to manually take pictures of the different plots in the experiment. Of course, we also took calibration pictures.
Usually, with the drone, we use the software Pix4D to calibrate our pictures taken by the Sequoia. In this case this is not possible, because the pictures lack geographic information (and probably also other information, I do not know about that exactly). I would now assume that I can simply calibrate these pictures using for example a raster calculator in a commonly used GIS. What I am not sure about, is the procedure underlying this calibration process. Could you maybe give me some advice where I could find the calibration information for the Sequoia camera? Or can you tell me how to best proceed?


#2

There is information available about this on this forum.
Roughly speaking you will have to do the following:

  • vignette correction
  • fisheye distortion correction
  • image registration (alignment)
  • radiometric calibration

It is not as simple as pulling up the images in GIS software and doing the radiometric calibration. You will have to preprocess the data before computing any vegetation (or research) index.

However, there is a software here that does all of that. You might wanna give it a try.


#3

Hi @maria,

Since you use Pix4D, they should be able to help you make Pix4D detect the calibration images even without GPS or select them manually through the user interface.

Be careful about not blocking the Sunshine sensor when acquiring data.
Please note that without the Sunshine sensor, it is not possible to perform consistent calibration except under constant solar irradiance (this is very rarely the case).