| Resumo: | The growing dependence of modern-day societies on electricity leads to the increasing importance of effective monitoring and maintenance of power lines. Due to the population’s renouncement to the installation of new electric power lines, the existing ones are constantly operating at maximum capacity. This leaves no room for breakdowns, as it leads to major economic losses for the electrical companies and blackouts for the consumers. Endowing Unmanned Aerial Vehicles (UAVs) with the appropriate sensors for inspection the power lines, the costs and risks associated with the traditional foot patrol and helicopter-based inspections can be reduced. However, this implies the development of algorithms to make the inspection process reliable and autonomous. Visual detection methods are usually applied to locate the power lines and their components. Although, they are generally too sensitive to atmospheric conditions and noisy background. Poor light conditions or a background rich in edges may compromise their results. In order to overcome those limitations, this dissertation addresses the problem of power line detection and modeling based on the use of a Light Detection And Ranging (LiDAR) sensor. A novel approach to the power line detection was developed, the Power Line LiDARbased Detection and Modeling (PL2DM). It is based in a scan-by-scan adaptive neighbor minimalist comparison for all the points in a point cloud. In the segmentation, the breaking cluster points are detected by an analysis of their planar properties. Exporting the potential power line points to a further step, it performs a scan based straight line detection. The final model of the power line is obtained by matching and grouping the several line segments detected using their collinearity properties. Horizontally, the power lines are modeled as a straight line, while vertically are approximated to a catenary curve. The algorithm was tested with a real dataset, showing promising results both in terms of outputs and processing time. From there, it was demonstrated that the proposed algorithm can be applied to real-time operations of the UAV, adding object-based perception capabilities for other layers of processing.
|
|---|