This is a drone methodology that we’re driven to explore ourselves at UAVWILD. Not only are Orangutans one of our bucket-list group of species to work with, but they also have a significant amount of human pressure upon them pushing them closer towards extinction.
Photo credit: Gunung Palung Orangutan Project.
For species in this position, it is increasingly crucial that the research conducted upon them is efficient and accurate. As we’ve mentioned before, low-cost drones address both of these elements, but they also allow us to undertake research aerially- which is a very useful tool for orangutan monitoring…
Something you may not have known is that orangutans sleep in nests made in trees (image below). They tend to make a new nest everyday as the move throughout the forest. Because of this behaviour, we can monitor the presence of orangutans by identifying these nests that are visible to drones.
Orangutan nest from drone imagery. Photo credit: Orangutan Survey.
An example of this method in action is the work done by Gunung Palung Orangutan Project in Borneo. Gunung Palung National Park is home to an estimated 2500 Bornean Orangutans, which is roughly 14% of the remaining wild population of this subspecies. The team at GPOP spent over a year in the field carrying out orangutan drone surveys, to compare this method to their traditional walking line transect methods on the ground- a time consuming and costly method.
Drone imagery orthomosiac, showing orangutan nests identified during the study. Photo credit: Gunung Palung Orangutan Project.
Their drone surveys collected roughly 10,000 images throughout the survey area. All of these images were processed by pure human power! Whenever an analyst came across a nest, they would circle it, and save a copy of the image. An orthomosiac (image above) was then made stitching together all of the drone images for each transect area, followed by the final nest count.
Although this would've been a long process, machine learning would make it a lot quicker after you've created the training data through this initial stage of manual work during the first round of surveying.
The research team commented “although the time investment in this project is long and the work ever so slightly tedious, our team are excited to see the results and are already hopeful that this method will indeed be a realistic tool for the long-term monitoring of orangutans across their entire range.”
If you’d like to take a further look into the work done by Gunung Palung Orangutan Project and get a more detailed view of their drone monitoring, follow the link above!
Kommentarer