Software
Autopilot Software
Using the APM Mission Planner, the flight paths of each mission can be pre-programmed by clicking on waypoints in a Google satellite map interface. The drone can be programed to take off and land autonomously, and circle over any waypoint for specified number of turns or duration. Operators can also program other flight parameters such as ground speed and altitude of each waypoint. Each pre-programmed mission is then uploaded to the drone, which will fly the mission autonomously upon launch.
Combining Aerial Videos with APM log
If you are using a GoPro HD camera, or any non-GPS enabled video camera, you can use a new web service, GPS4Sport.com, which allows for the synchronization of your APM log (in GPX format) with your video (must be uploaded to youtube or other video-hosting websites). GPS4Sport then displays an animation of your flight in Google Earth in your browser with an insert showing the synchronized video footage. Playback requires Google Earth Plug-in to be installed in your browser. This service is free for now,and relatively user friendly.
See below for an example of a Google Earth-synchronized video of our orangutan nest hunt in Sumatra.
Geotagging Aerial Photos using APM log
Although we use a GPS-enabled camera for our drone missions (Canon SX230HS), the GPS data on this camera might not be as precise as those logged by the APM. Fortunately, the Mission Planner contains a hidden function for geotagging aerial photographs using GPS data logged by the APM. Here I describe how to use this function.
- Select the aerial photographs you wish to geotagged and put them in a new folder.
- Also put the “.txt” APM log file for that same mission in the same folder. (Note that this “.txt” log file is not from the telemetry. You’d have to download it from the APM to your laptop via the Terminal screen of the Mission Planner.)
- On the “Flight Data” page of the Mission Planner, hit Ctrl-F. This brings up a window with several hidden functions.
- Click on “Geo ref images”. This brings up a new window for this function. See image above.
- Use the button to browse to the “.txt” log file. Also browse to the Directory (folder) where the pictures are located. If your Canon camera was on GPS time during the mission, then there would be no need to specify a time offset. Otherwise to calculate time offset between GPS time and camera time, take a picture of a GPS-corrected clock with your camera (e.g., time display on a handheld GPS). For example, if your camera time is 5 seconds behind GPS time, then you should enter “-5″ in the time offset box.
- Hit the “Do it” button. This creates a series of 5 new files in the picture folder, including a KML, a GPX, and three text files containing the new 3D coordinates of each picture. For example, this is what you’d see in the file “location.tel”.

- To actually stamp the APM’s coordinate info onto the EXIF data of your photographs, you need to click on “GeoTag Images”. This creates a duplicate set of pictures, each with a suffix “_geotag”, containing the new 3D coordinates from the APM log.
Post Processing
The most straightforward form of post processing of aerial photographs is to stitch them together using commercially available software, such as the AutoPano Giga. The resultant mosaic is essentially a near real-time land use/ cover map that could be used for detecting and monitoring land use change.
This is an 18-photo composite image of Zurich Zoo’s new elephant enclosure. Aerial photographs were taken by a camera on board an autonomous Conservation Drone 2.0 that was flying at a speed of 10 m/s at 100 m above ground. The camera (Canon SX230HS) was CHDK-programmed to take photographs at 3 second intervals for 5 minutes (shutter speed 1/800, ISO 100, focal length 5.0 mm). **Upon request from Zurich Zoo, I have reduced the resolution of this image on Flickr to 1024 x 813 pixels, which is <10% of the original size and resolution.[/caption]
DroneMapper.com
We have also teamed up with DroneMapper.com, which operates an innovative SaaS system which allows drone pilots to manage their flight data, build beautiful geo-referenced orthomosaic maps, DEMs and 3D models from aerial imagery.
Documentation of camera set-up by Dronemapper.com:
- Detailed CHDK firmware installation instructions (Link)
- Detailed graphical instructions on optimal CHDK settings (Link)
- Guidelines on aerial data collection and flight planning (Link)
3D forest model of a transect in Sumatra, Indonesia, based on aerial images acquired by Conservation Drones.
Data processing by Dronemapper.com. Animation by Ervine Lin of ETH Future Cities Laboratory (futurecities.ethz.ch).

DroneMapper-processed orthorectified mosaic from one of Conservation Drones’ Orangutan Nest Hunt transects.

DroneMapper-processed 3D surface model from one of Conservation Drones’ Orangutan Nest Hunt transects.

DroneMapper-processed 3D surface model with ortho overlay from one of Conservation Drones’ Orangutan Nest Hunt transects.
Disclaimer: The Conservation Drones team (Lian Pin Koh and Serge Wich) strives to develop Conservation Drones that are operationally stable and safe. However, we have no control over the operation of Conservation Drones by our collaborators. And therefore, the Conservation Drones Team will be not held liable for any bodily harm and/or property damage resulting from the operation of Conservation Drones by third parties, including our collaborators.















![JPG [Group 1]-Sandajaya300M 110_Sandajaya300M 513_dehazed-172 images JPG [Group 1]-Sandajaya300M 110_Sandajaya300M 513_dehazed-172 images](http://farm9.staticflickr.com/8512/8420149544_255756e68f_m.jpg)



Wonderful imagery product. I noticed a linear feature misalignment in the above image. Have you managed to work that bug out of the system yet? The commonly accepted spec on misalignment is 2-5m post-processed.
What software do you use to create the elevation models
Hi William, the Mission Planner software grabs elevation data from Google Earth’s DEM.
Hello. I was wondering if their is any open source software that is capable of creating georeferenced ortho-mosaics of aerial images from UAVs as well as DSM generation?
People have been using Hugin (http://hugin.sourceforge.net/) and then georeference the image using any open source GIS software or commercially available packages such as ARCGIS. I am not aware of any easy solutions to the DEM part, but there are certainly people working on this. For instance see: http://www.andrewhazelden.com/blog/2012/03/exploring-aerial-photogrammetry-using-bundler-and-meshlab/
I hope this helps. Serge
Thanks for the tips Serge! I work for an organisation carrying out restoration work on degrading peat bog habitats in the uplands of the UK. We’ve been looking to develop a UAV capability to collect aerial imagery and then generate DEM’s to help us determine where best to target restoration work as to further research of these precious habitats. Can you provide help and advice in setting this up? I’ve been looking at some of the commercially available options but they are completely out of out budget unfortunately but I can see so much potential for high resolution data like this.