A novel tree trunk recognition approach for forestry harvesting robot

Robotisation of forestry harvesting in New Zealand has the potential to achieve great productivity benefits and support the timber industry in the face of global market competition. Recognition and localization of tree trunks is the first critical operation for an autonomous forestry harvesting robot and is a challenging task due to variations of illumination under normal forestry harvesting conditions. This paper presents a novel method of recognizing tree trunks based on Hough Transforms of the L*a*b* colour space representation of the harvesting scene. The test results show that the proposed algorithm is able to correctly determine the location and shape information of tree trunks in various lighting conditions. This method lays a strong foundation for further research on autonomous operation of forestry harvesting robots in steep terrain.