Paper presented at the 16th European-African Regional Conference of the ISTVS
Last updated
Title: Obstacle detection vision system enabling autonomous mounding on clearcuts
Authors: Songyu Li, Håkan Lideskog, Magnus Karlberg, and Ruben van Westendorp
Abstract: After final felling, a clearcut needs to be prepared as soon as possible for subsequent planting with the purpose of restoring forest reserves. In Sweden, the multi-row continuously advancing mounder is a mature and widely used solution for site preparation. It is driven by a base machine, where rotating mattock wheels make upside-down mounds on the area to prepare for subsequent planting. During operation, a plurality of obstacles such as stumps, residual trunks, single trees, large stones, and the like, remain after felling and are scattered at the site, forcing the operator to continuously watch out for them. These obstacles can to some extent be avoided through control of each mattock wheel arm, which causes inconvenience to the operator and reduces the efficiency of the mounding procedure. Thus, by adding automatic obstacle avoidance functionality to the mounder, the efficiency and working environment can be significantly improved. In this paper, we present a feasible solution for real-time detection of obstacles in front of the mattock wheels of a mounder using a realistic simulated forest environment combined with a vision-based deep learning detection algorithm. The developed system can effectively identify most target obstacles and provide a reference for the operation of the mounder, either manually or by machine control. The successful implementation of this system also provides a starting point to further improve the automation level of mechanized site preparation, while further improving the mounders’ working efficiency.
Order the full paper:
ISTVS members: receive three papers per year as part of your membership via the ISTVS Member Portal: