Fitting oriented bounding box algorithm for accurate positioning of transformer wiring terminals
Yuntao Zhao, Jie Huang, Weigang Li · 2023
As an important connection component of the power grid, transformer wiring terminals need regular maintenance and repair. Some problems exist in traditional maintenance such as high labor cost, high operation risks, and low efficiency. Relying on the point cloud processing technique, this paper proposes a F-OBB (fitting oriented bounding box) algorithm based on surface normal estimation and applies it to the high-precision positioning of terminal blocks. Surface regions of the rectangular-like terminal are separated by the clustering method of region growing after filtering pre-processing. Then the optimal normal vector of each region is solved by the spherical coordinate system transformation and least-squares method to determine a bounding box to represent the target and obtain the positioning parameters (position and orientation). The practical test results show that the proposed method can generate bounding box highly close to the target, accurately extract the positioning parameters and keeps good robustness in various scenarios.