CleanSat - SIG AI
Reference number | |
Coordinator | Stift Skogsbrukets Forskningsinstitut Skogfor - STIFT SKOGSBRUKETS FORSKNINGSINSTITUT, SKOGFORSK |
Funding from Vinnova | SEK 6 398 189 |
Project duration | November 2019 - January 2023 |
Status | Completed |
Venture | AI - Leading and innovation |
Call | From AI-research to innovation |
Important results from the project
The project aimed to develop, test and implement an AI model to predict the need for pre-commercial thinning in forests based on remote sensing and logging of the operators´ work. A model has been developed and validated in one forest area with promising results. Preparations for implementation are underway and there is great interest from both intended users and IT suppliers. Challenges linked to data collection and modeling have meant that some of the high goals had to be modified during the course of the project.
Expected long term effects
The project has resulted in a model for estimating the need for clearing at pixel level, based on remote sensing, data from the clearing law and AI methodology (neural networks). The model has not yet been fully implemented, but the results and the validation carried out within the project have created great interest in the industry and made it clear which parts need to be further developed in order to get a functioning decision support for estimating clearing needs. Further work is planned within the framework of the Mistra Digital Forest research program.
Approach and implementation
The project was a collaboration between two research organizations (Skogforsk and Örebro University), three forestry companies (Mellanskog, Sveaskog and Södra) and one service company (Field). The constellation made it possible to collect and process data connected to real clearing objects, build AI models, validate the results in the field with subject specialists and prepare for implementation at a service company as well as at the companies themselves. Challenges consisted of securing data for modeling and evaluation as well as staff turnover during the project period.