Weakly supervised power line detection algorithm using a recursive noisy label update with refined broken line segments

• Published in: Expert systems with applications Vol. 165; p. 113895
• Main Authors: Choi, Hyeyeon, Koo, Gyogwon, Kim, Bum Jun, Kim, Sang Woo
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 01.03.2021; Elsevier BV
• Subjects: Accident prevention; Algorithms; Annotations; Artificial neural networks; Datasets; Image segmentation; Industrial application; Labeling; Labels; Line segments; Machine learning; Pixels; Power lines; Semantic segmentation; Training; Unmanned aerial vehicles; Weakly supervised learning; Line segments; Semantic segmentation; Weakly supervised learning; Industrial application; Power lines
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2020.113895

•A weakly supervised power line detection algorithm is proposed.•A line specialized refinement algorithm is applied for performance improvement.•A repeated learning framework with noise label develops the model output. Detection of power lines in aerial images is an important problem to prevent accidents of unmanned aerial vehicles operating at low altitudes in the electrical industry. Recently, pixel-level power line detection using deep learning has been studied but production of the pixel-level annotations for massive dataset is difficult. In this study, we propose a power line detection algorithm using weakly supervised learning method to reduce the labeling cost for dataset generation. The algorithm is divided into two stages. First, an approximately localized mask was generated based on a convolutional neural network which was trained with only patch-level labels. Second, recursive training of segmentation network with refined broken line segments was executed. A refinement algorithm, line segment connecting (LSC) is a power-line-specialized refinement module that connects broken lines by approximating the segments as partially straight. In proposed algorithm, predicted image at each recursive step was updated as a label of the next training and the label was developed by itself with LSC. The comprehensive experimental results of our algorithm showed state-of-art F1-score of 94.3% in weakly supervised learning approaches on public dataset. This result suggests that the proposed algorithm is useful for low labeling cost with high performance in line detection application.