FORESEE

The sustainable provision of wood as a renewable raw material makes an important contribution to the environment and society as well as to sustainable economic development. This is especially true for rural areas. It is not only forest fires and dead trees that endanger this raw material; in northern Europe, winter storms also pose a serious threat. A major problem in forest protection is the lack of continuous recording of information about the forest and its analysis in well-founded risk models.

The project developed measurement systems and evaluation processes for monitoring and inventorying forest areas in order to ensure the sustainable supply of wood as a renewable resource. The core of the overall system is a lightweight drone scanner that is specially equipped for measuring and segmenting vegetation thanks to its multispectral imaging. Based on this, a data processing pipeline was developed.

The project focused on the following areas:

Firstly, the focus was on developing a new multispectral LiDAR system so that it can be used both on a civilian drone and as a backpack solution. For the first time, geometric data sets with additional spectral information can be generated. Furthermore, the new system combines data sets collected within the forest (ground-based) and from drones. This unique approach enables a completely new analysis of the forest and thus actively contributes to assessing the health status of individual trees with unprecedented spatial and temporal resolution.

A second focus was the high-resolution and GIS-based risk and resilience analysis for the forest management system. With the help of the high-resolution and data-based condition data of the forest, the risk of unintentional tree felling due to storms can be predicted. In addition, the most efficient time for planned tree felling and management can be determined, taking into account all natural and economic aspects.

The third focus was on the development of an AI-based selection process for laminated wood production. For the selection of laminated wood, a database with sufficient characteristics and correlated mechanical parameters was initially created for one type of wood. For this purpose, automated image analyses were used, which were supported by AI methods (e.g. neural networks) and correlated by corresponding mechanical tests for test specimens with selected characteristics.