From Harvesting Machines to Field Sensors: Assessing the Accuracy of Tree Stump Positioning Recorded by Cut-to-Length Harvesters

Modern forest harvesting operations are increasingly performed using highly mechanized cut-to-length systems, particularly harvesters and forwarders. These machines are characterized not only by high operational productivity and improved operator safety, but also by the growing integration of advanced onboard computer systems. Such systems enable the collection, storage, and transfer of detailed information on machine performance, harvested timber, and stand conditions within forest machine fleet management platforms.

As a result, modern harvesters are no longer only tools for timber harvesting. They can also function as mobile data acquisition platforms, recording spatial and operational information during harvesting operations. One of the potentially valuable datasets generated during machine work is the digital location of felled trees or their remaining stumps. The accuracy of such spatial data is important for forest management, post-harvest verification, operational planning, and the further development of digital forestry solutions.

The aim of this study was to assess the positional accuracy with which forest harvesters determine and record the location of tree stumps in a digital spatial layer. The study was based on data collected from three harvesters operating in three forest districts managed by the Regional Directorate of State Forests in Olsztyn, north-eastern Poland. The analysed machines included two Komatsu 901 harvesters and one Komatsu 931 harvester manufactured by Komatsu Forest AB.

For each machine, three sample plots were established. Within these plots, the positions of tree stumps were measured manually using RTK-GPS equipment and then compared with the stump positions recorded automatically by the harvesters. The obtained datasets were analysed to determine the magnitude of positional deviations between manual field measurements and machine-recorded locations, as well as to assess whether statistically significant differences occurred between machines, plots, or measurement conditions.