It has been a crazy 48 hours since first arriving in Costa Rica. I haven't yet had the time or internet access to post any updates, so I will try to outline the main events from the last couple days. We have had exciting results thus far and have greatly exceeded our expectations for the experiment.
We are experimenting with the future of turtle research and have the possibility to impact how sea turtle research is gathered and used throughout the entire world.
As mentioned in my introduction post (Sea Turtle Introduction), I have partnered with two international conservation/research companies and am spending the week experimenting with a variety of drones and equipment. The general premise of our experiments rely upon taking pictures of nesting sea turtles as they lay their eggs on the beach. When they do this, the turtles drag themselves across the sand leaving a large trail across the beach (0.5-1.5 meters wide, see picture at right). Currently, researchers patrol the beaches searching for tracks and then they can document and put tracking devices on any turtles they may find. For the most part, these turtles only lay their eggs at night (between 8pm and 4am) so that they can be gone before sunrise. Therefore, the majority of the turtle patrols are done routinely in 30 minute intervals throughout the night.
Night #1 - Infrared Drone Tests
We arrived at our basecamp (located approx. 30 minutes from the primary research beach) in the late afternoon on January 6th. Being that we arrived after sunset, we decided to proceed with the Infrared Drone Tests objectives along the beach. Using a modified camera from which I removed the visible spectrum lens and coupled with an IR illuminator, we are able to see in complete darkness. Humans and sea turtles are unable to see the infrared wavelengths, making it an ideal spectrum for nighttime research.
After about one minute of our first sweep we immediately encountered an endangered Leatherback turtle laying her eggs. In fact, there was a second Leatherback turtle further down the beach.
"in a known breeding area there was over 1500 LEATHERBACKS in the 1990's. today that area has seen a 98% decline with less than 40 know leatherbacks now."
We were fortunate enough to see two of these extremely rare and endangered Leatherbacks on our first night alone. Their wide bodies provided perfect tracks for testing out the IR camera system. Below is a picture from the camera and short video of the IR flight. Although obviously difficult to see, considering it is pitch black and we are using inexpensive consumer grade sensors, we had remarkable results:
In addition to using the drones to capture aerial pictures/videos, we brought along an IR ground camera to capture the turtles on the ground. Being that one of the Leatherback turtles laid her eggs too close to the water's edge and near the main entrance, we had to relocate them into a new hole we dug further down the beach. Below is a short video with our footage from the night:
Day #1: Autonomous drone tests
After spending most of the night running patrols on the beach and infrared tests, we set aside the morning for conducting our primary objective of daytime aerial imaging. Ultimately, we were hoping to conduct a fully autonomous mission of mile long beach. This means that a researcher or myself would be able to walk out to the beach, turn the drone on, and the computer automatically takes off, flies, photographs, and lands itself. In order to accomplish this, I established a checklist of smaller objectives to first execute:
- Fly drone manually and ensure GPS signal strength (Check)
- Conduct autonomous flight from waypoint to waypoint without camera (Check)
- Conduct short flight from waypoint to waypoint with camera (Check)
- Determine best altitude for photographing the entire beach (Check)
- Test long-range autonomous flight of the entire beach (Aborted, Check)
- Acquire properly overlapped images for georeferencing (TBD)
- Georeference and stitch together imagery for 2D/3D models (TBD)
After running through these initial objectives it was obvious that the drone was operating much better than anticipated. The autonomous flight modes perfectly executed each mission and was able to cover the entire length of the beach. When we first tested the long range flight, we lost visual contact with the drone in the tree line, so I aborted the mission and blindly flew the drone back to the home location. After reviewing the logs, it appears everything was operating perfectly, we were just unable to see the drone at the distance of a kilometer.
After a few minor adjustments we were able to view the pictures from the drone and had some amazing results:
These pictures were both taken at an altitude of 50 meters, more than twice the maximum height we anticipated being able to fly. As you can see in each image, it is clear where the turtle crawled up the beach. The top image contains the track of a Black Turtle (the straight thin lines were a guide for us). The bottom image contains a day old track from a Leatherback Turtle, as seen in our IR flights. Overall, these images possess far greater quality than anticipated, allowing us to even view individual footprints of people.
After finishing some of these tests I had some time to film at an adjacent beach. Hopefully I will be able to upload more/high quality footage soon:
The next step of the project is to determine the optimal height, speed, and overlap for the drone to fly at. I plan on finishing these tests in the morning and shortly thereafter posting a full update that goes in depth on the drone flight and mapping software we will be using.