On this page

Multi-dimensional filtering and intelligent extraction algorithm for structural parameters of cable point cloud images

By: Zihan Dong 1, Wenchao Ding 1, Hong Wang 2, Wangqiang Wu 1, Lei Lei 2,3, Liang Wang 2,3
1State Grid Weinan Power Supply Company, Weinan, Shaanxi, 714000, China
2State Grid (Xi’an) Environmental Technology Center Co., Ltd., Xi’an, Shaanxi, 710000, China
3Electric Power Research Institute of State Grid Shaanxi Electric Power Co., Ltd., Xi’an, Shaanxi, 710000, China

Abstract

This study first introduces the basic principles of laser triangulation and the principles of handheld laser scanners, and then applies point cloud reduction and point cloud smoothing methods for data processing. Addressing the issue that conventional point cloud filtering methods can cause data degradation in noisy environments, this study proposes a point cloud filtering algorithm that combines dual tensor voting and multi-scale normal vector estimation. By comparing different filtering algorithms and conducting visual analysis across various scenarios, the proposed method is evaluated. Additionally, the Q3D software is used to establish a cable layout model and perform simulation calculations. Experimental results show that the improved algorithm demonstrates good robustness in different scenarios, effectively enhancing noise removal rates while minimizing the loss of environmental features, and maintaining good algorithmic efficiency. Additionally, simulation results indicate that the improved method can quickly and conveniently extract cable distribution parameters. Finally, by applying the proposed method, an improved cable design scheme is proposed, and it is found that the average stranding pitch of the improved power cables is 167, within the standard range.