Powerline Feature Detection and Extraction Using Photon Lidar Point Clouds

Date

2019-06-04

Journal Title

Journal ISSN

Volume Title

Publisher

item.page.doi

Abstract

Remote sensing technologies are a critical component for managing utility infrastructure. Utility companies rely on GIS data for planning, engineering design, operations, severe weather response, vegetation management, asset management, disaster management, and many other usages. LiDAR offers cost effective and scalable solutions to many of the utility use cases. Due to the advances in its technology, Photon LiDAR can capture data at a much faster rate and from a higher elevation than linear-mode LiDAR. However, high point density of Photon LiDAR point clouds presents a challenge for feature detection and extraction. This research presents an integrated approach for effective and efficient, detection and extraction of powerline features from highly dense 3D point clouds. The method involves selectively filtering the point clouds, detecting linear features using Hough transform, and extracting LiDAR points for electric features using RANSAC algorithms. Once the candidate points have been extracted, machine learning algorithms are employed to classify the points into specified target classes, namely, conductors (wires) and poles, with high levels of accuracy, precision and recall.

Description

Keywords

Powerline ampacity, Electric utilities--Management, Optical radar

item.page.sponsorship

Rights

Citation