From: mk_thisisit
Artificial intelligence (AI), particularly machine learning and computer vision, is being increasingly applied to assist in wildlife conservation efforts. These technologies enable rapid analysis of vast datasets, helping scientists and forest guards monitor and protect endangered species and their habitats [01:34:00].
Emaza Application
Emaza is an application developed by Apsilon, a Polish company, in collaboration with scientists from the University of Stirling in Scotland and national parks in Gabon [01:16:00], [09:57:00]. It was recognized by an international institute operating under UNESCO, being placed among the top 10 “outstanding” applications globally for its contribution to UN Sustainable Development Goals, specifically the preservation of biodiversity and protection of life on land [00:40:00], [00:52:00], [07:06:00], [07:32:00].
Functionality
The application primarily aids in saving endangered species in Central African forests, which can be the size of Poland [00:00:00], [01:05:00], [02:12:00]. It utilizes machine learning algorithms trained to recognize species in photos taken by camera traps [01:39:00], [01:43:00]. These camera traps are small devices with heat and motion sensors, triggering a photo when an animal, such as an African forest elephant, passes by [02:20:00], [02:30:00].
Ecologists and forest guards place these traps during expeditions, then retrieve SD cards containing tens of thousands of photos, for example, 30,000 photos from a single expedition [02:07:00], [03:02:00], [03:19:00]. Manually reviewing these photos, which are often complex due to dense forest imagery, can take two to three weeks, leading to decreased efficiency and accuracy over time [03:35:00], [03:42:00].
Emaza automates this process:
- It operates offline on a laptop, which is crucial as research bases deep in forests often lack internet access [03:24:00], [04:17:00].
- A batch of 30,000 photos can be processed in about 8 hours (one day), significantly reducing the time from 15 days (three weeks) [04:36:00], [04:42:00].
- The accuracy of the AI models is comparable to manual human review, with errors balancing out in ecological models [08:59:00], [09:16:00].
Applications in Conservation
By quickly processing data, Emaza enables scientists to:
- Track animal migration: Assess how migration patterns (e.g., of elephants) are changing due to factors like climate change, providing insights that were previously unavailable [05:07:00], [05:12:00], [05:25:00]. This information can inform decisions, such as avoiding road construction in animal migration routes [05:51:00].
- Monitor disease spread: Quickly identify gorillas in photos to allow experts to assess their health and understand disease spread, aiding in community protection and intervention [05:59:00], [06:06:00].
Availability and Funding
Emaza is open source, with its full source code and machine learning models available on GitHub [06:27:00], [10:02:00]. This ensures the international community, including scientists and national park guards, can utilize and further develop the application for nature conservation [06:42:00], [10:06:00]. While some small funding has come from non-governmental and governmental institutions, the project has largely been a pro bono initiative by Apsilon as part of their “data for Gold” program [10:13:00], [11:40:00].
The application is currently used in national parks in Gabon, Cameroon, and Kenya, and has recently been tested in Congo [00:19:00], [06:56:00].
Broader AI Applications in Conservation
Beyond Emaza, Artificial Intelligence is increasingly vital for sustainable development and planetary well-being [12:04:00], [12:21:00]. Projects globally are leveraging machine learning to tackle various environmental challenges. For instance:
- In cooperation with the Arctowski station of the Polish Academy of Sciences in Antarctica, AI is used to detect animals from drone images over islands near Antarctica [00:30:00], [13:01:00].
- At the North Pole, AI models accelerate the large-scale study of plankton, which is critical as it serves as food for ocean life and significantly impacts climate [13:21:00], [13:27:00].
These examples highlight how AI is becoming an indispensable tool for scientists needing to quickly analyze large and complex environmental data sets, fostering a deeper understanding of ecosystems and informing effective conservation strategies [12:49:00].