DJI drones create thermal images (infrared) in a thermogram format, which is not compatible with current photogrammetric pipelines (e.g., Agisoft Metashape, Pix4D). Moreover, the DJI thermal data should first be alibrated according to emissivity, relative humidity and camera-target-distance. This procedure enables one to convert entire folders from .rtpeg to .tif format while doing the calibration on the fly. The procedure utilizes the DJI Thermal SDK. This repository is a fork that has been optimized for parallel processing to greatly shorten the run time.
Examples of single image frames acquired with the DJI h20t.
Example of an orthophoto generated from the .tif files created with this procedure.
Important: Do not worry if the default image viewer on your OS will show 'white' output *.tifs. The standard image viewers are not designed to resolve images with float data format [0-1]. Just proceed with the output with your photogrammetry software (e.g. Agisoft Metashape). You may want to inspect individual pixel values in a GIS or advanced image viewer software (Irfan view, ImageJ).
For Metashape-Users: You may want to adjust the brightness and contrast settings in Metashape for visualizing the individual images. For most datasets setting the brightness to 1 percent and contrast to 200 percent works fine.
Originally created by Teja Kattenborn and modified by me.
If you want, you can cite the repo using the provided citation file (see above).