Artificial Intelligence

An artificial intelligence system to preserve ecosystems

An artificial intelligence system to preserve ecosystems

Chile has 42 National Parks, cataloged as areas established for the protection and conservation of flora and fauna. Precisely the challenge of conserving these territories from possible threats, a research team of Natural Resources Engineering of the Universidad del Bío-Bío developed a proposal based on AI.

The proposal makes use of cameras already placed in the wild areas, the artificial intelligence system identifies and qualifies the exotic species, typical of the area. In addition, this new tool allows the automation of the species classification process, saving 99% of the time used by the professionals of the National Forestry Corporation of Chile (CONAF) for this task.

"The way we deliver artificial intelligence to the user is through a very simple application, and the AI simply processes the photograph. We have already trained the AI with the database of Chile's National Parks, so it is very well designed to solve the problem of invasive species," said José Díaz, a graduate of Natural Resources Engineering and member of the research team.

As a data base and artificial intelligence training, the team chose Cerro Castillo National Park, located in southern Chile and listed as one of the most outstanding tourist destinations to visit in 2022, especially recognizing the work done by the Agriculture and Livestock Service, CONAF and the Rewilding Chile Foundation in the conservation of the huemul, a species in a state of vulnerability and part of the national coat of arms, as it is considered a Natural Monument of Chile. Currently, the team is already working on a second version of the application that will be delivered to CONAF rangers in other territories.

The application is worked in open-source license, which allows its dissemination and future collaborations to continue developing it and improve its potential to other instances.

Bastían Díaz, 4th year student of Natural Resources, commented "We are solving a national and global challenge that is climate change and we are focused on providing a solution to nature. In addition, we are working on the application with open source because we want to open the knowledge or what we create and grow on a larger scale".

The system was developed within the framework of the Climate Change Hackathon, one of the strategies of the South-Sub-Antarctic Consortium Ci2030, a formative instance that challenges researchers to develop innovative scientific solutions that solve global problems with a local perspective. The Consortium is composed of six regional universities: Universidad de Talca, Universidad del Bío-Bío, Universidad Católica de la Santísima Concepción, Universidad de La Frontera, Universidad Austral de Chile and Universidad de Magallanes.

Read more at: https://ciencia2030.cl/2022/11/09/estudiantes-consorcio-ciencia-2030-desarrollan-inteligencia-artificial-para-clasificar-especies-invasoras-en-parques-nacionales/

20 de Diciembre, 2022



metodika