Wild Intelligence Lab - WIL
Land Classification – a Machine Learning Use Case
Inspired by the research done within the SAVMAP Project, we use machine learning models to classify the landcover into different soil, grass, and sand types. These data allow conservation professionals to accurately determine and predict the amount of food sources available to animals, as well as taking the right precautions to prevent severe animal population loss. In addition to the extraction of valuable data, the models are used to aid annotation for more complex deep learning techniques.
Animal Detection – a Deep Learning Use Case
Modern Instance Segmentation models such as Mask R-CNN allow the pixel-wise classification of objects in images. To apply those models to the high-resolution drone images, we explore the use of different architectures and inspect their performance. The model output will allow detection of animals, vegetation, and even aardvark holes. In addition to the classification, the detection also enables the extraction of key data such as the size of the animal or the area of a tree’s crown.
The latest from Wild Intelligence Lab
Table of Contents Introduction In Africa, one of the majestic Big 5 is facing extinction due to the high pressure of illegal poaching and habitat
Table of Contents Project Introduction As giraffe have been listed as ‘Vulnerable’ by the IUCN Red List of Endangered Species and their numbers across Africa
Drone Adventures is a non-profit organization which was founded in Switzerland in 2013. It flies drone missions throughout the world, connecting people who know drones best with people who need them the most. Drone Adventures’ goal is to demonstrate and promote the great potential of drones to protect our planet and support local communities.