Detection of steel materials and bolts from point-clouds of power transmission pylon

Since power transmission pylons have long life cycles, they have to be repeatedly maintained by replacing deformed and corroded steel materials. In conventional maintenance, three or more workers have to climb on a pylon and measure the dimensions and bolt positions of damaged steel materials in order to manufacture replacement steel materials. However, such works are dangerous and costly. In this paper, we discuss methods for precisely calculating dimensions and bolt positions of steel materials using dense point-clouds captured by a terrestrial laser scanner. Since steel materials consist of planar surfaces, plane detection methods are useful for detecting steel materials. In our method, steel materials are detected by combining the RANSAC method, the thinning of planar regions, and the region growing method. We also propose a bolt detection method by fitting points to bolt shapes. We evaluated our method using actual point-clouds of a power transmission tower. We also evaluated the accuracy using sample steel materials. Our experimental results showed that our method achieved sufficiently good accuracy for practical use. © 2020 CAD Solutions, LLC.