My organization has been using our Autel EVO Ⅱ Dual 640T drone, equipped with a FLIR Long Wave Infrared (LWIR) sensor to map temperature across stream environments. This is being used to assess heating and cooling inputs as habitat for native salmonids. We have had some success processing data using Pix4d in areas with lower density riparian vegetation but in areas with more dense canopy the images are not calibrating. I think in part due to the low resolution of the infrared photos and in part due to the fact that heavily treed areas are difficult for the software to create tie points with. I am curious if others have had better success using Agisoft Metashape? The file types I have been using is the .TIFF files as our drone only exports .jpeg and not .rjpeg unfortunately.
26 May 2025 5:08pm
Not the same equipment, but at work I have used Metashape with DJI drones with an IR sensor (Mavic 3 Thermal, Matrice 3DT, Matrice 4T, etc) and the workflow to create IR maps relies on metashape fromt TIFFs created using the DJI thermal SDK. There are a lot of considerations that will affect the final quality of the output:
- Flight speed should be low to avoid motion blur, the IR sensors are quite sensitive to that.
- high overlap is required for more homogenous ground features
- the processing settings in metashape play a key role - to avoid the "fishbowl" effect etc.
- Flight altitudes need to be consistent to avoid calibration issues.
I think it is worth giving it a shot in Metashape. From my experience, metashape tends to error out much less frequently. I reckon the forest scene might be the biggest challenge for you - given the low number of tie points etc. What are your flight settings ?
Nihal Mohan Moodbidri