Laser scanning from harvesters - a smart way to detect sweep and crook of trees during felling?

Laser scanning sensors are becoming cheaper and can increase precision in forestry. However, no one has tested whether the technology can be used to assess crook or sweep of trees during harvesting - a decision support system with this capability would increase the resource efficiency of the forest industry. This is currently being explored within Mistra Digital Forest, in collaboration with Komatsu, who would like to develop the technology.

In Sweden, trees are quality-assessed at the time of felling and are cut into logs with a specific use in mind. The advantage of this is that the log can go directly to the "appropriate" industry, but it also places great demands on the decisions made in the forest. One Achilles' heel in this approach is the sweep and crookedness of the trees. Sweep is difficult to estimate with the naked eye, and because it is difficult to take it into account when the tree is cut, much of the value of the forest can be lost.

- Five, and sometimes up to ten per cent of the logs in a pine stand are classified as wreckage, that is, they are sorted as " second-rate", and of these, more than half have too much sweep. This is a waste of resources and shows that a better basis for decision-making is needed, says Nils Lindgren, researcher in value chains at Skogforsk, and he continues:

- How the logs are cut is determined by a digital decision support system that optimises the value of the trunk, based on information about tree species, length and diameter. By collecting laser data and developing systems to assess the sweep and crook of the trees, the decision support system gains access to information that makes it better at optimising the value of the forest. Cutting and sorting the logs incorrectly and sending a log with excessive sweep to the sawmill, for example, also means unnecessary diversions and transports.

Ever more accessible and attractive technology

The fact that the market for sensors is growing and the technology is becoming cheaper makes it increasingly justifiable to use laser scanning to detect sweep and crook already in the forest. In this project within Mistra Digital Forest, the researchers have installed a sensor on a harvester; the sensor belongs to the segment that is interesting from a cost perspective, and that can be assumed to have a sufficiently high degree of detail to detect sweep and crook. A baseline has been collected using a stationary laser scanner with a very high degree of detail, in addition by manually sampling trees on some of the felling area. To test the potential of different types of mobile data collection, selected parts of the forest are also laser scanned by drone as well as manually by a person holding a scanner.

This first step brings together researchers and industry stakeholders to develop a prototype. Achieving sufficient accuracy in sweep measurement will not be easy, says Nils Lindgren, adding that another major challenge is building a system that works in the harvester's exposed environment. This is one of the challenges that Komatsu is focusing on through its involvement:

- Regarding the actual technological development, developing a sensor that is robust is a challenge. It has to withstand harsh weather, vibrations, and it must be able to clean away water droplets that interfere with data collection, for example, says Mattias Nyström, Senior Design Engineer at Komatsu Forest.

An important first step - but it's only the beginning

Much of the work is also realated to with making the right choices in a jungle of different laser technologies.

- Many new stakeholders have appeared recently, and we need to find the partnerships that have longevity over the next ten years. We have to develop a solution that works flawlessly. Measuring sweep is a highly prioritized first step, but we see many other interesting applications linked to laser scanning from harvesters, says Mattias Nyström.

New field tests are planned for the summer of 2024, and when the project ends in 2025, we hope to have laid the foundation for the development of applicable decision support tools.

- By the time our work has progressed so far that it has resulted in concrete decision support, the price of sensors will probably be at a level that makes this a fully realisable technology, says Nils Lindgren.