The consumer models of some drones do not display their position, altitude, camera heading, and camera declanation during flight, but do store this data in the EXIF and XMP metadata of their still image captures
It is strongly suggested that the user should calibrate the drone's magnetometer (compass) before each flight. The image metadata from an un-calibrated drone can be a few degrees off from the correct heading. This can result in dramatic Target-resolution inaccuracies. Always verify a target match location from OpenAthena before use
E.x.:
parseImage.py has an experimental feature for automatic extraction and use of EXIF/XMP sensor information from drone photos
parseImage.py can be run in an interactive mode, or headless mode based on how many arguments are given after python parseImage.py. Interactive is designed for use by humans with one or more images, headless is designed for use by automation with one or more images. The output of headless mode is subject to change in future versions
Before using parseImage.py, make sure you have a valid geoTiff file that covers at least the location of the drone and the possible location of its subject. One way to do this is to get the drone's location from an image's metadata, then clip a geoTiff file of the surrounding area
Next, choose an image for which to resolve the location of its subject. For example DJI_0419.JPG:
Finally, run parseImage.py. To save time, the geoTiff file can be provided as the first argument after python parseImage.py (or python3 parseImage.py on Macs for example):
python parseImage.py cobb.tif(Note: the file must end in .tif. If you do not provide a geoTiff file, you will be prompted for input of a filename)
RETURN
Enter the full file path of the drone image if it is not in the src directory. Otherwise, just input the image filename and press RETURN
Multiple images can be processed at once in interactive mode. When you're finished inputing images, type done to begin processing
Let's copy that NATO MGRS into Google Maps:
Pretty good!
(note: the NATO 10-digit grid without GZD is underlined for easy reference. We need to include the full MGRS with GZD to give the precise location. An alternate, Warsaw pact style SK42 Gauss-Krüger grid ref is similarly underlined)
If parseImage.py is given any additional arguments after a geoTiff file, it will be run in headless mode. It will try to read each argument (past the geoTiff file) as an image filename. It will try to resolve a target and put a resolved location in a file IMAGENAME.JPG.ATHENA for each image.
Image parsing can fail for any image for a variety of reasons, including the search running out of bounds, location not covered by the geoTiff, etc. However, an IMAGENAME.JPG.ATHENA file will only be generated where both the image metadata extraction and the terrain search process were sucessful
In a future version, these files of extention .ATHENA will be replaced by .gpx and be of the GPS Exchange Format
A Docker image for parseImage.py is available here
The openathena Docker image is convenient because it comes with all pre-requisites and openathena in a ready-to-go container that can be run on demand
For running with your own GeoTIFF DEM and drone images, mount a directory on your host system your desired files to a directory inside the /home/user in the container. Then pass as arguments a GeoTIFF file and one or more drone images using relative file paths like ./recon-photos/Altis.tif and ./recon-photos/DJI_0666.JPG
It is highly recommended to choose a particular tag from the releases page rather than using the latest tag whenever openathena is to be used in a automated production environment
E.g (with Docker already installed):
docker pull t3l3tubie/openathena:v0.1.26-pre-alpha
docker run -v ~/Desktop/recon-photos/:/home/user/recon-photos openathena:v0.1.26-pre-alpha ./recon-photos/Altis.tif ./recon-photos/DJI_0666.JPGThe container will run and perform its task, then exit. A file like ~/Desktop/recon-photos/DJI_0666.JPG.ATHENA will be available on the host filesystem, containing the data of the target calculated by OpenAthena
One tool to view EXIF and XMP metadata on Mac, Linux, and Windows is Phil Harvey's (not affiliated with this project) exiftool
An equivalent command line tool available on Mac and Linux is exiv2 (also not affiliated)
The following is an example of a command to show the sensor metadata of an image taken from a particular DJI drone:
exiv2 -P kt DJI_1234.JPG | grep -i "gimbal\|latitude\|longitude\|alt"drone-dji.GimbalYawDegree is the direction of the aircraft's camera, from -180 (exclusive) to +180 (inclusive). 0 degrees is North, 90 is East, -90 is West, etc. Internally this number is converted to an azimuth for use by OpenAthena
drone-dji.GimbalPitchDegree is the amount of pitch of the camera's gimbal. While the DJI displays this number as negative, Internally the absolute value of this number theta is used for OpenAthena. Values of below -90 (straight down) produce undefined behavior and are prohibitted from being entered during manual data input