Key Point Grasping Method for Wire Core Oriented to Mechanized Secondary Wiring of Power Grid

View at publisher Summarise with VIRAT
Authors: Xiaofan Song, Wei Qian, Qimin Li, Ling Zu, Congying Wu, 瑞 谭
Cited by: 0
References: 15

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

To achieve the mechanized transformation of secondary wiring operations in power grids, aiming at the difficulty that the S-shaped wire cores in wiring operations are not conducive to visual perception and robotic grasping due to their slender structure and single color, a complete robotic arm grasping technology ranging from image perception to grasping point extraction is proposed. In the visual perception stage, the open-vocabulary detector Grounding Dino is integrated with the visual model SAM. Meanwhile, to address the interference of backgrounds similar to wire cores, the CLAHE algorithm is adopted as preprocessing to enhance images. In the key point estimation and extraction stage, a skeletonization-based method for centerline extraction is designed, and a neighboring point sorting algorithm is proposed to optimize centerline generation. On this basis, scale-invariant grasping key point estimation is realized based on the principle of similar triangles, and a 5-DoF planar grasping pose for collaborative robotic arms is generated. Physical verification is carried out on the built wiring robot platform. Experimental results show that the average success rate of the proposed grasping method reaches 91%. Compared with traditional learning-based methods, the proposed method can effectively improve image perception capability and antiinterference performance in the perception phase, realize zeroshot localization and accurate mask extraction of wire cores, and generate stable grasping poses.

Fields of Study & Subjects

Robot Manipulation and Learning Power Line Inspection Robots Advanced Neural Network Applications

Publication details

Journal / source
International Conference on Electrical Automation and Artificial Intelligence (ICEAAI)
Pages
593–599
Publication date
16 January 2026
Language
EN