Reflectance by Phantom4 Multispectral

Hi Support,
Values in reflectance maps became extreme small (0 – 0.01) when using band images taken by DJI Phantom4 Multispectral. I think it usually take values between 0 and 1. Image Exif contains “Sun Sensor” value and I selected “Sun Irradiance” for the radiometric corrections. Does “Irradiance Gain” in Exif effect? How can I calculate correct reflectance normalized by irradiance?

Hi,

We will be in contact with you in the internal ticket. I am replying to this for the community: Using radiometric targets will help you get absolute reflectance values from the DJI P4 Multispectral. Please do not leave out the reflectance factors as 0.

Kind regards,

Hi, I have same problem with Phantom 4 Multispectral data

Hi there,
I have got a similar question as I´d like to know how the program calculates the reflectance value using the metadata.
There is a digital number in the Xmp.Camera.SunSensor Tag and I have found a formula to calculate the physical radiance from it. But I don´t understand how the scalar for the radiometric calibration is calculated (I assume there is a scalar between 0 and 1 and every pixel of the raw image is multiplied with that scalar to get the reflectance value. Is this correct?)
The explanation for the sunsensor values says:“Xmp.Camera.SunSensor contains the raw value (in digital counts), one scalar for each channel of the image.”
I don´t understand the explanation. In my case this raw value is 2650. How can it be a scalar?
greetings, Timo

Hi,

This camera/drone has a sunshine sensor, thus radiometric correction will be done. However, unless you are using radiometric targets, the reflectance maps generated will give you relative values and not absolute reflectance values that’s why you see small values. But you can get the NDVI because it is a self normalized index. You can use any panel as soon as you know the particular wavelength of that camera. For MicaSense panels, the reflectance for any band is automatically detected.

However, if you are using the same camera for your study, you can do a temporal study or even a spatial comparison. However, the camera/drone has to be the same. You can have a look here in the reflectance target section: https://support.pix4d.com/hc/en-us/articles/202559509-Radiometric-corrections

Dear Fernanda,

I know about reflectance maps either absolute and relative reflectance. But, I have a question here: I have in hands parrot calibration target (picture) that I was using with Parrot Sequoia. But now, I am using Phantom 4 Pro Multispectral. So, this calibration target can only be used with Parrot Sequoia or Can I use with P4 Multi. Detail is, P4 Multi has blue band neither avaiable in Sequoia. If not, Will I need to calibrartion this target in lad to get correct albedo values?
Thanks

Hi,

The reflectance factor values will not be the same. This is because the bandwidth and band wavelength will be different for both the sequoia camera (the sequoia target reflectance factor values you see is for that particular wavelength) and DJI p4m. You can reach out to parrot to see if you get the reflectance factor values for all wavelengths (it will be a graph or a table).

This is the band wavelength and width for sequoia:

This is for DJI P4m:

Blue (B): 450 nm ± 16 nm; Green (G): 560 nm ± 16 nm; Red (R): 650 nm ± 16 nm; Red edge (RE): 730 nm ± 16 nm; Near-infrared (NIR): 840 nm ± 26 nm

Dear Momtanu,
Thank´s for your help. I understand that wavelenght and bandwith area different for each sensor. Becaouse of this, I wll need to do do panel calilbration with spectrodariometer (laboratorry off course). I did not understand so much Fernanda´s information about relative and absolute reflectance. These valeus are have not so epxressive differences. I have used Parrot Sequoia, and remember values were not differents unlles used relfectance target.
By the way, you told to contact Parrot to get panel information. It is impossible because Parrot support does not answer questions for anyone.
Regard´s

Could you calibrate in your lab and get the reflectance factor values? That will also do as you said. If you are using a radiometric target, you will get absolute reflectance values and the values will range from 0-1. Otherwise, if you use only DLS sensor, you will get relative reflectance which means you will not be able to compare outputs/index map/reflectance map from another sequoia camera or any other multispectral camera or suppose results from a journal paper/article… However, if it is from your own camera (the same one), you can compare outputs.

Hi all,
Let’s assume we get the correct factor values (from parrot or lab calibration) for using the parrot sequoia calibration target in our DJI P4 MSP processing. Can you explain how this can be done automatically, without the need from us manually to type in the reflectance factor value in the radiometric calibration window? At the moment the calibration target is automatically selected and also the light blue rectangle is drawn (so it is detected) but the factor values are missing.

Hello Bert,

I hope you are doing great. If you do that, and if you can share this with us will be great. But I can not promise that Pix4Dmapper will populate those values automatically. I can suggest it to the team, of course. However, you can always add the values manually in the dialogue, as you mentioned.

About the measurements with the Parrot target I consulted with our especialist and this are his recommendations in case you would like to do it:

“The Parrot target does not report reflectance values for the P4M bands. You need to measure them yourself. First, measure the spectral reflectance factor with a spectro-radiometer. Then you need to know the spectral sensitivity of the P4M bands, and using that you can compute the effective nominal reflectance value of the Parrot target for the P4M bands.”

Sincerely,

Hi Fernanda,

Could you please check with your development team and get back to us on whether the factor values can be populated automatically using a sequoia or any other calibration target (we could also try with a micasense calibration target)? Is there anything we can do on our side for automatically populating the factor values? Because it is practically impossible for each flight to fill in manually those values as we do a lot of flights during the season to monitor plant growth, and the season is starting soon. It would be great if Pix4d could support this.

Kind regards

Hello,

Thanks for the patience. So Pix4Dmapper will not automatically give the reflectance factor. What you can do is to reach out to MicaSense and ask if they can provide you with those values. For the wavelengths of the P4M:

Blue (B): 450 nm ± 16 nm; Green (G): 560 nm ± 16 nm; Red (R): 650 nm ± 16 nm; Red edge (RE): 730 nm ± 16 nm; Near-infrared (NIR): 840 nm ± 26 nm

Just in case if you are using Pix4Dfields, then the software will automatically recognize and use the images of your MicaSense panel.

Hope this helps.

Sincerely,
Fernanda

Hi Fernanda,

Apologies if my question wasn’t very clear.
Lets assume we already have the correct reflectance factor values.
Is there a way to store this information somehow so we don’t need to insert these values manually each time we do a processing with a DJI MSP flight?
Thanks a lot for your response and help!

Kind regards,
Christina

Hi Christina,

Thanks so much for your patience. If you have the correct values, then they have to be inserted manually.

Hope this information helps you.

Best regards,
Fernanda

Hello, if I understood correctly, it is possible to use the Micasense panel with phantom 4 multispectral images in Pix4Dfields and the software will automatically recognize the panel and work the absolut calibration by simply loading the panel images together with the images taken with the P4M. Is that correct?
Otherwise, if I don’t use the panel but always get my datsets with the P4M, could I still compare the vegetation indexes obtained in diferent days and/or weather conditions?
Thanks

Hola una consulta:

Pero las imagenes corregidas radiometricamente solo con el sensor de Luz del P4 Multiespectral, si pueden ser comparados en vuelos de distintas fechas? por ejemplo que haga un NDVI hoy y otro la semana entrante a diferentes horas del día?

Gracias

Eddison

Hola Edison,
Esto depende del sensor en sí. No podemos hacer una afirmación informada en cuanto a la comparación de los resultados entre diferentes fechas con esta cámara. Mi recomendación es que confirmes esto con DJI sobre su rendimiento radiométrico o fiabilidad sin un panel de reflectancia. Por nuestra parte, hacemos todas las correcciones posibles con la información proporcionada.

M;uchas gracias

Eddison

Hi any updates if Pix4D can now process multispectral images from DJI P4 Multi-spectral drone?