Tree Root Detection and Cutting Control of a Pipeline Operation Robot Using Vision and Force Sensing in Sewer Pipelines
Year: 2025
Type: conference-paper
Abstract: Traditional manual inspection and maintenance methods for urban drainage pipelines suffer from low efficiency and high safety risks. This paper proposes a vision and force sensing-based autonomous tree roots detection and cutting control method for a pipeline operation robot in sewer pipelines. The robot is equipped with a depth camera and utilizes the YOLOv11 model to identify pipeline defects, achieving target recognition. The robot is also equipped with a 5-degree-of-freedom robotic arm. We established the kinematic and dynamic models of the robotic arm, designed a trajectory planning method, and verified its feasibility in a simulation environment. Finally, we conducted experimental studies. The results show that the robot system achieves a recognition rate of over 60% for tree roots, with depth accuracy at the millimeter level. The angular control deviation of the robotic arm motor is within 1°. This research provides a foundation for tree root cutting and other types of obstacle detection and cutting in underwater pipelines.
Authors: Yuanwen Zheng, Xiatian Zhou, Wenze Cao, Jiaxuan Kang, Xin Chen, Jun Zhang
Institution: Southeast University
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4411665286