Community Forum

Author Topic: Calibration questions on Dji Phantom 4  (Read 1723 times)

uplift

  • Newbie
  • *
  • Posts: 1
    • View Profile
Calibration questions on Dji Phantom 4
« on: September 14, 2016, 10:48:14 PM »
Hello.
I am new to calibration, but i have bought Agisoft Pro, and i am going to do measurements on piles of gravel and soil with Dji Phantom 4. So as beginner to this, I have some questions:

1: Do i need to do a calibration and import it to Agisoft, or does Agisfoft sort it out by itself?

2. (if so) how do i import the file in to Agisoft?

3: I did a calibration, does all the needed parameters show? -and is the tangential and decenting distortion the same?  (see attachment)

4. The results i got when pressing OK, was this, does it looks normal(okay ? See picture: PHANTOM 4 RESULTS (3)

Thanks a lot in advance, as a new man to this, its a bit of a head-scratcher.
I really appreciate your time and help.

Kind regards, Anders.
« Last Edit: September 14, 2016, 10:54:48 PM by uplift »

Yoann Courtois

  • Full Member
  • ***
  • Posts: 230
  • Engineer in Geodesy, Cartography and Surveying
    • View Profile
    • ALTEA Experts ( Lyon, FRANCE )
Re: Calibration questions on Dji Phantom 4
« Reply #1 on: April 06, 2017, 09:35:12 AM »
Hi Anders !

Don't know if you'll see this answer somehow, so I'll not make a long and detailled one :)

I have thinking about calibrate our DJI Phantom 4 Pro UAV before handle surveys. But there is a huge problem:
The calibration should have been made with exactly the same parameters than the survey will !!!
So,  as you have to always take pictures (your whole survey) with exactly the same focus (basically with UAVs, focus on infinity for lots of surveys), you have to calibrate your camera while focusing on infinity... ::)

And, with a simple calculation of depth of field, you quickly understand your calibration has to be made more than 10-15m far from your chessboard...which is ensuing a huuuge chessboard :o

As a conclusion, I would say your calibration will be almost useless as it looks you took pictures only a few meters far from the chessboard, and maybe in autofocus.

And there is the moment I may need help: How to calibrate that kind of cameras ?
I've thinking about handling a georeferenced survey of an ideal place (ideal for matching) with autocalibration, and then iterate the alignment while using calculated camera parameters...
Have somebody even handle that ? Or have another solution to calibrate cameras focusing on infinity ?

Thanks :D
-------------------------------------------
           Yoann COURTOIS
Engineer in Geodesy, Cartography & Surveying ( graduated from INSA Strasbourg, FRANCE )
Topographic engineer in ALTEA Experts, special. in Photogrammetric survey methods ( Lyon, FRANCE )
-------------------------------------------

Anaz

  • Newbie
  • *
  • Posts: 2
    • View Profile
Re: Calibration questions on Dji Phantom 4
« Reply #2 on: May 08, 2017, 12:15:26 PM »
Hi yoan,

Me too,
I have no idea to calibrate my p4pro correctly.
Hoping someone find it & share the method here.

Anaz


barrubba

  • Newbie
  • *
  • Posts: 34
    • View Profile
Re: Calibration questions on Dji Phantom 4
« Reply #3 on: May 08, 2017, 12:28:47 PM »
Using agisoft lens?

Yoann Courtois

  • Full Member
  • ***
  • Posts: 230
  • Engineer in Geodesy, Cartography and Surveying
    • View Profile
    • ALTEA Experts ( Lyon, FRANCE )
Re: Calibration questions on Dji Phantom 4
« Reply #4 on: May 09, 2017, 10:43:06 AM »
Hi barrubba !

Yes basically we could calculate a accurate calibration using Agisoft Lens...

But there is still the problem of the chess board... The calibration has to be made with the exactly same parameters as the survey will ! So as I explained before, at 15m or more, far from a huge chessboard.

And even though a solution can be found... I've noticed that the P4Pro doesn't "save" the camera parameters when turned off. So each time we start the UAV, the camera is set up as default (focal distance for example). So I easily imagined we cannot get the exactly same focus between two flights...
So... would it be impossible to handle very accurate measurements with that kind of device ??? ?
-------------------------------------------
           Yoann COURTOIS
Engineer in Geodesy, Cartography & Surveying ( graduated from INSA Strasbourg, FRANCE )
Topographic engineer in ALTEA Experts, special. in Photogrammetric survey methods ( Lyon, FRANCE )
-------------------------------------------