Hello,
I am trying to create a fast orthomosaic using WebODM. Unfortunately, the images I am using are for research and cannot be shared. Here is my problem statement:
I am trying to create a fast orthomosaic of thermal images output by a DJI M30T camera. The images output by this camera are not absolute temperatures, they are just a ‘heatmap’ single-channel JPG. However, absolute temperature can be acquired by running these JPGs through the DJI SDK, which can temperature correct the images and then output them. However, once they are output, they obviously cannot be the typical image format we use (uint 0-255), they are instead an array of float32s (to allow for temperatures like 15.6° to be described).
So far, I have used ODM to
- Stitch the RGB JPGs into a fast orthophoto for this survey
- Stitch the UNCORRECTED thermal imagery into a fast orthophoto for this survey
and both worked fine. However, stitching the temperature-corrected (float32 .TIF) images into a fast orthophoto processes and completes as normal, but the output I download results in a very low resolution (~10x less than what is output when using the uncorrected thermal imagery) orthophoto. Additionally, the projection seems wrong, as loading it into GIS software places it flipped and in the wrong spot.
Any idea why the temperature-corrected float32 images, when used in stitching, result in such a lower resolution?
Thanks!