Hackathon to save the forests!

Hack for Sweden 365 hackathon took place online during the weekends of the 21st & 22nd of November and the 5th & 6th of December, 2020. The hackathon was organised by DIGG - the Agency for Digital Government and the challenge was provided by the Swedish Space Agency with Space Data Lab, the Swedish Forest Agency and the Swedish University of Agricultural Sciences (SLU) with the Forest Data Lab: how to solve the spruce bark beetle problem that causes mass death to the spruce forests.

More than half of Sweden is covered with forests, with pine and spruce being the most common trees. If you ever considered how much damage the spruce bark beetle does to the forests and spruce trees; imagine loading all damaged trees from 2019 on timber trucks and putting the trucks as close as you can to each other – they would reach all the way from Malm? to Kiruna and back again, approximately 3500km! The economic loss is huge as the value of the tree is only 20% of a healthy tree.

Kim Hedberg, Hampus Hellstr?m, Yurdaer Dalkic and I competed in the hackathon and worked on a solution for an open data platform. Our solution was awarded as the most innovative and the one using open data in the best way at the first hackathon weekend. We were also selected as one of three finalists to compete at the second hackathon weekend.

With use of available satellite images, maps, weather and forest data, we built an AI powered solution that will notify forest owners of the areas with the highest risk of being infested as well as when the warming season is about to begin and it’s time to start searching and removing infested trees. Satellite images are only able to identify changes in large forest areas over the time so in order to identify infected trees people often use drones. The solution also included a web, as well as a mobile application to create a user-friendly interface making it possible for forest owners to consume the knowledge both at home and in the forest.

We conducted several interviews forest owners and learned that a marketplace for 3rd party services such as drone operators or sniffing dogs would be helpful to increase the usage of the methods that are considered the most effective to identify infested trees. We also trained a Custom Vision AI model to be able to recognise trees from the provided images. The accuracy of the model could be further improved over the time with additional pictures taken from drones.

With the use of weather data and algorithms based on how the spruce bark beetle develops from egg to insects, we were able to create a model that predicts when the spruce bark beetle is active and by sending mail and push notifications to the forest owners, they can act to find infested trees as soon as possible, reducing the spread. In the video we made you can see main features of the application.

The interviews with forest owners also revealed that a large amount of time is spent looking for damaged trees. By offering a solution for hikers and hunters who spend a lot of time in the forest where can report damages that they see, the forest owner would get valuable help and save time. Our solution therefore has the possibility to take a picture of the damage and report it in the app. The app uses the geolocation data from the image, maps it with the owner of the forest and notifies the owner.

We had a lot of fun at this hackathon, trying out some new concepts and learning a lot about forests and spruce bark beetle problems. Even if our solution was not selected as the winning solution in the finals, we strongly believe there is a lot of potential in the ideas and the suggested solution that could make a difference for the forest industry. We are hopeful that we will one day see such open data platform where all users, regardless if the forest owner is independent or a large forest company, can join forces in the battle against the spruce bark beetles and other forest threats.

Thanks for reading, feel free to like or share!