With the help of Artificial intelligence (AI), the Swedish Forest Agency will develop a solution that identifies larch trees damaged by the insect larch sack template. The goal is for the solution to be used in the long term to deal with other forest damage. The project is made possible partly through Microsoft's AI for Earth program, where the Swedish Forest Agency as the first organization in Sweden received grants through the program.
The Swedish Forest Agency is to develop a solution that, with artificial intelligence, identifies the degree of damage of larch sack molds and helps to prevent the attacks being dealt with earlier. Going forward, the Swedish Forest Agency hopes to also be able to use the knowledge in larger projects and use the same scheme to deal with other forest damage. In a press release , Gunnar Isacsson, ecologist at the Swedish Forest Agency, says that the next step is to identify alms-affected elms in Gotland.
- The Swedish Forest Agency has initiated an incredibly important project, not only by identifying and contributing to measures that promote nature. The project also contributes to building a new knowledge base that also enables other countries to build on the results, says Daniel Akenine, chief technology and security manager at Microsoft Sweden.
Both data from the project and the finished solution should be made available to users outside Sweden. The project aims to contribute to an increased global understanding of the damage caused by the larch sack molds.
Artificial intelligence and drone images behind a new solution
Sweden's most severe damage to the larch sack is currently in southeast Västergötland. The Forest Board's project uses around 1,600 drone images across that area to make more than 40,000 markings that will teach the model to identify all larch trees and classify them as healthy or damaged. By quickly finding which trees are most stressed and where in the terrain the attacks are greatest, the forest owner can more easily find and remove the weakest trees. This can also minimize the economic damage. In areas in the United States and Canada where larch sack is invasive, infestations in new areas can be identified early and the spread is counteracted.
- Environment and nature are sometimes the most complex systems that exist and if we are to preserve our nature, we need to understand our ecosystems better and what affects them. Today we have new advanced tools for collecting data and processing it using sensors, machine learning and cloud services. Technology that didn't exist just a couple of years ago. It makes me hopeful for the future, says Daniel Akenine.
As a consequence of climate change, the damage caused by larch sack mills is feared to increase in Sweden. In Sweden, the larch bag mill is found in Götaland, Svealand and along the Norrland coast. Usually, the insect pestle larch sackmeal is aggressive in those parts of Europe that have a warmer climate.
For more information about the project and expert opinions, take part in the Forest Board's press release .
About the project
The project runs until February 2020 and is part of the National Forest Data Lab, which is a meeting place for analyzing and making available forest data. Other actors in the project are the Jakova Innovation Center and the B3 Consulting Group, which works with data annotation and the AI model. The Forest Board uses Microsoft's cloud service Azure to build the AI model and as part of AI for Earth.
About Microsoft AI for Earth
Microsoft AI for Earth supports projects that use AI to promote and change how people and organizations manage the earth's natural systems and resources. To date, 435 grants have been awarded to projects with an impact in over 71 countries. Read more about AI for Earth and how to apply for a grant here .
Illustration: The Forest Agency. Using drone images, tree-level markings are made to teach the model to identify all larch trees and classify them as healthy or damaged.
For more information contact:
Daniel Akenine, Technology and Security Manager at Microsoft Sweden, Daniel.Akenine@microsoft.com
Halil Radogoshi, System Developer, Forest Agency, email@example.com