Skip to contents

DJI cameras

DJI_cameras.Rmd

In this article we will discuss the workflow if you just want to use the Thermal SDK functionallity of DJI.

1. Some background

The thermal images from DJI cameras are stored in their own specific way. For a start they are written as JPG files instead of tiff. Furthermore, the DN values in these thermal images cannot be readily converted to temperatures. This conversion is limited to the DJI software. However, DJI released a Thermal SDK with Dynamic Link Libraries which can be incorporated into your code to process DJI thermal images. On big remark: these can only be accessed on a Windows or Linux system… In this R package the Thermal SDK is built in as well. To access the libraries, the package runs a c++ function (get_temp_dirp_cpp()) in the background using Rcpp.

2. Create a ThermalUAV

If your have a DJI camera that is compatible with this package (check through tuav_cameras()), you have two options: - Follow the image-based workflow - Process the data with the DJI IR Processing tool (dirp)

In the first case I gladly refer to the article about the image-based workflow. In this article we discuss the second option. The DJI IR processing tool is built into the function tuav_dji(). This function processes the data using the provided libraries from DJI. The input variables are thus the same as in their DJI Thermal Analysis Tool software. To keep the geotags and facilitate the process, first a ThermalUAV object should be created before using the tuav_dji().

thermal_uav_dji_M3T <- tuav_create(path = "E:/Thermal_Project/Thermal_data_dji/DJI_20240806165425_0078_T.JPG",
                               # ThermalUAV objects can also consist of only 1 image if desired
                               camera = "DJI_M3T",
                               meta_csv = NA,
                               flight_height = NA)

3. Process with DJI IR Processing tool

Once this ThermalUAV object is created, you can call the function tuav_dji(). Please check carefully the range in which each parameter should be provided.

dji_M3T <- tuav_dji(thermal_uav_dji_M3T,
                    obj_dist = 25,      # limited to [1-25] in meter
                    rel_hum = 47.2,     # should be between [20 - 100] in %
                    emissivity = 0.985, # Should be between [0.1 - 1]
                    refl_temp = 5,      # Should be between [-40.0 to 500.0 deg C]
                    return = TRUE,
                    export = FALSE,
                    export_path = NA)

When processed, you can plot your image:

library(terra)
LST_dji_M3T <- rast(dji_M3T)
plot(LST_dji_M3T, main = "LST", col = map.pal("magma"))