Thermal Orthomosaic with FLIR DUO PRO R 336

[denis1807]

Hi,
I have purchased a new FLIR DUO PRO R 336 thermal camera + RGB camera, now I want to build an thermal orthomosaic with that and I don't know where's the correct procedure.
The camera produce rjpeg image that contains both thermal (640x512 px image) and RGB ( 4000x3000 px) with al the radiometric data inside.

The image is produced with drone at 40m with a verttical and horizontal overlap of 75%.
Please, how I have to process everything???
can anyone describe the procedure step by step???
I think it would intersting for many.
Thanks a lot.
Best regards.

 

[denis1807]

sorry nobody who can kindly help me???
At this point I'm interested to know if I can or not elaborate Thermal orthophoto with Agisoft
Thanks

[paulm]

Have you looked at Chapter 3 of the manual (http://www.agisoft.com/pdf/photoscan-pro_1_4_en.pdf)? That should provide at least some general information on how to approach your task...

[denis1807]

Thanks for your reply, but I know perfectly how Agisoft works with normal and multispectral datasets, but the problem is how it works with Thermal image, because I'm not able to obtain a correct results.

Best regards

[denis1807]

Ok, I'm able to build a Thermal orthophoto with a new dataset of image with different paramenter.
Now I want ask if is possible to export the radiometric data in some way.
Thanks

[James]

Did you convert the thermal images to tiff format in flir software, or are you using the jpgs with the embedded radiometric data?

my understanding is that the radiometric data inside the jpgs is only available to flir software as it is not stored in a way that any other software would know how to read it, so non flir software just treats it as a normal jpg and the radiometric data is never seen.

however, i suspect that if you use the appropriate flir software to convert to tiff files (i think flir researcher or something like that) then photoscan will be able to access the radiometric data stored therein and also export in the same format, which may then also be readable by flir software.

i do have access to an old chunky flir camera but nothing much in the way of software so can't try working with tiff files unfortunately.

[denis1807]

Hi James,
thank you for your answer.
Your observation is correct, I have understood the same with the rjpg problem in external software.
The only problem I have find in Flir Tools is the impossiblity to convert rjpg files to tiff, but from tiff I can convert them to rjpg.
Therefore I will have to re-test with a new tiff dataset.
then I'll let you know the result.
Hello and thanks.
have a nice day

[James]

Does your camera have the option to save directly to tiff?

I don't seem to have that on our old P45 but would be so useful.

[denis1807]

Yes the DUO PRO R can save in RJPG or TIFF.

[grant]

I have exactly the same question.  In theory, the solution seems straightforward:

1) Use SfM software to construct an orthomosaic from the visible images;

2) Copy the computed orientation/position data from each of the constituent visible images to the corresponding thermal images;

3) Using those orientation/position data, construct the thermal orthomosaic, with averaging of radiometric values in areas of overlap.

The question is whether software tools that can do this automatically have already been created and, if not, whether there's a non-automated way to achieve the same result.

I would think the folks at FLIR would have an interest in providing such a tool, since it would greatly increase the usefulness of the Duo Pro R for certain applications.

[jmgc]

I've opened a new topic on this subject here: https://www.agisoft.com/forum/index.php?topic=13049.0

Thanks