Open AccessResearch on 3D Reconstruction Technology of Power Line Based on Image Semantic Segmentation
Author(s) -
Xionggang Li,
Feng Zhang,
Guoqiang Li,
Dongwei Xiang,
Chengcheng Yang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2095/1/012015
Subject(s) - line (geometry) , artificial intelligence , computer science , power (physics) , computer vision , noise (video) , segmentation , pixel , process (computing) , catenary , image segmentation , image (mathematics) , mathematics , geometry , physics , quantum mechanics , operating system
Because it was difficult to distinguish the characteristics of the power lines by the traditional methods of extracting the power lines, which led to the current situation of incomplete reconstruction and a large number of noise in the process of rebuilding the power lines only by the inclined photographing. In this paper, the power line information in the image is segmented pixel by pixel by introducing in-depth learning semantics segmenting neural network. The three-dimensional coordinates of the power line are calculated by the principle of multi-view three-dimensional reconstruction. Finally, the power line is fitted by the catenary equation to complete the incomplete power line reconstruction. The results show that the fitted power line model has high accuracy and meets the requirements of power related applications. Based on the traditional three-dimensional reconstruction, a new idea for power line reconstruction is proposed.