Home // IARIA Congress 2025, The 2025 IARIA Annual Congress on Frontiers in Science, Technology, Services, and Applications // View article

RGB-D Object Classification System for Overhead Power Line Maintenance

Authors:
José Mário Nishihara
Heitor Lopes
Thiago Silva
André Lazzaretti
Andre de Oliveira
Ronnier Rohrich

Keywords: RealSense; Machine Learning; Object Classification; Transmission Lines; Autonomous Inspection.

Abstract:
This paper presents the development and evaluation of different machine-learning models applied to classify objects in high-voltage transmission lines using depth data captured by a RealSense D415 camera. Four models, k-Nearest Neighbors (kNN), Decision Tree, Neural Network, and AdaBoost, were tested using simulated and real data collected in a laboratory environment. The results show that the kNN and Neural Network models achieved robust performance, while the Decision Tree model faced significant limitations due to excessive nodes. Moreover, tests with real data revealed noise in the images, which affected model performance. This study also highlights the feasibility of using depth cameras for autonomous inspection tasks, potentially reducing costs and enhancing safety in high-voltage environments.

Pages: 55 to 60

Copyright: Copyright (c) IARIA, 2025

Publication date: July 6, 2025

Published in: conference

ISBN: 978-1-68558-284-5

Location: Venice, Italy

Dates: from July 6, 2025 to July 10, 2025