Hi
I am just getting started with Pix4D and I have been trying to purchase a radiometric calibration target with know albedo values to use for camera calibration. I have visited a number of photography shops and they have no idea what I am talking about. I have done search online and only found Pix4d or scholarly articles about radiometric targets. Can you recommend a place to buy a target or the manufacturers. I am located in Vancouver, Canada.
Many Thanks
Hello Haydn,
In regards with the calibration target, you should search for a white balance card (you could find plenty of them in Amazon). The problem with these cards is that they normally give the Albedo values for Red, Green, Blue. When they say for example 18% Gray Card, this means that they have Albedo values for Red (0.18), Green (0.18), Blue (0.18).
When you have a camera with different bands, then you should know which is the reflection of the material at each specific band.
This is a more complicated task. You should better check with the manufacturer of these cards. From our side, we are trying to find which is the reflection of the most common cards in different bands. Once we have more information, we will make it public.
We have found, for example, the following company that sells more complete calibration targets (they come with the albedo values in infrared too):
http://multispek.com/products/calibration-targets?variant=1033650412
However, they are much more expensive.
Regards,
Thank you this information is very helpful. I look forward to hearing the results of you investigations
Regards
It seems like the terms ‘albedo’ and ‘reflectance’ are being used interchangeably. I have access to a spectroradiometer that can measure reflectance across the visible and near-infrared spectrum. I have some white teflon calibration targets with unknown ‘albedo’ values. If I measure the spectral reflectance of these targets and extracted the values for each band, would these be appropriate to use in the Pix4D calibration method?
Additionally, would it be necessary to measure the reflectance of the target each time I fly to account for changing light conditions? I am assuming not because the actual targets are not changing themselves and the reflectance or albedo is just a percentage of the light being reflected.
For the calibration image of the target used in Pix4D, does it need to be from the same altitude as the rest of the imagery or can it be taken with the UAV hovering a few metres above the target?
Duncan,
Albedo and reflectance are not the same as you probably suspect. Reflectance is a function of wavelength and albedo is not. So your spectroradiometer measures reflectance and that will vary with wavelength. What you need to do ideally is to resample the spectroradiometer data to the same band passes as your camera. Good luck getting spectral profiles of your camera, these may be hard to find and you may have to go to the manufacturer. Your spectroradiometer probably has band widths of a few nanometers and your camera more like tens of nanometers at least.
Also, you DO have to measure your calibration target each time. Getting your data to reflectance involves removing the effects of scattered light ,etc. and because that changes all the time, you need to measure your calibration target frequently - perhaps before and after a survey.
You do not need to measure the calibration target at the same altitude as your imagery. What is more important is to get lots of pixels covering your target so you get good stats to work with.
Google Labsphere for a source for your calibration panels.
Joe Zamudio
VP Remote Sensing UASUSA
Thanks for the info, Cody/Joe. I checked out Labsphere and then I remembered that we actually have a Spectralon target that came with our spectroradiometer. I guess the best thing for us to do moving forward would be to use that to calibrate the reflectance. My understanding is that Spectralon is a near perfect reflector (95-99% Albedo across 350-2500 nm range).
Dear all,
Then, having the reflectance, you could calculate the albedo values. The albedo values take into account the reflectance, the incoming light and the sensitivity of your camera.
Please note that you should have a lambertian reflectance surface (https://en.wikipedia.org/wiki/Lambertian_reflectance) as a target.
You do not need to measure the albedo values every time you fly (they are fixed values).
You should not take a picture of your radiometric calibration target from the flight height. It is better to go close to your target and take a picture by hand. As Cody said, “What is more important is to get lots of pixels covering your target so you get good stats to work with”
Best regards,
Will this be necessary with the Sequoia as it has a sunshine sensor?
Thanks
Rob
Dear Rob,
In theory, if you use the sunshine sensor, there is no need for a calibration target. However, the integration of this data is planned for 2.2, which is planned to be released close to the shipping of Sequoia. You can find more information about Sequoia and its features in our new FAQ on the subject:
https://www.pix4d.com/sequoia-faq/
Best regards,
Dear Rob,
At the moment, we still recommend using a calibration target if you have one since it provides more information for the radiometric corrections. Using only the sunshine sensor, the values can be compared across flights, but the reflectance values will be on a relative scale.
Best regards,
Hi,
Can you please elaborate on the ‘reflectance values relative scale’ mentioned above?
Are all bands of the sequoia cam being multiplied with (a different each time) constant or does each band have its own scale?
Is there any way I can obtain correct reflectance values from flights already performed with only the irradiance sensor and no calibration panel?
Thanks,
Dimitris
+1
Hi
I want to ask,why do we use the teflon calibration tag,always have the problem of overexposed?
Tip are as follow
Unable to compute radiometric target calibration for Sequoia_4.0_1280*960(green)with …Not enough valid pixels for band 0
Hi Xinyang,
If the digital number of the calibration target images is higher than 63000 the software will identify it as being overexposed. To reduce the likelihood of this happening, it is recommended taking a few snapshots of the reflectance target while on the field, and to lay the target on a more reflective surface ( light asphalt for example).
Cheers,
Ina
Hi Ina,
I just checked the images we got using different white balance cards, the values are all higher than 63000, e.g. 65472. I will try again tomorrow using the method you said.
Thank you very much.
Xinyang
Dear All,
I’ve tried to calculate sequoia irradiance myself through equations show in here
especially in SEQ AN 01 - Pixel to Irradiance.pdf 155
and then in order to calculate reflection of each band , how could I to do with or without calibrate panel ?
Thank you very much
Echo
Hi Yaoyao,
From our side, we do not have such methods or algorithms that could help you manually compute the reflectance.
I suggest you post your questions here:
https://forum.developer.parrot.com/t/reflectance-estimation/5597/2
Best,
Ina
Hi,
I did not find an albedo value for thermal reflection or the way to calculate! Is there a close matching value for a thermal reflection? The thermal sensor used is a FLIR Dou Pro R, 640x512.
Thanks.
Zaki
Hi Zaki, The thermal sensors having R as the suffix are radiometrically calibrated. You do not need a target. The temperature values in the generated Pix4D reflectance map are absolute temperature values. You could read some publications related to thermal calibration. Some people use ice or cold water close to 0 degrees C (keep that in a bowl in the field) so that it can be used as a reference.