A Tree Barrier Distance Measurement Method based on the Image Semantic Segmentation of Overhead Transmission Lines

Rapidly growing trees in proximity to overhead electrical lines present a notable risk to the consistent delivery of safe and reliable electricity services by utility companies. A key element influencing the security of these power lines is the proximity of trees. Yet, the swift identification of tree obstructions against complex image backgrounds and the accurate gauging of the tree-line distance remains a formidable challenge. This paper introduces a novel approach for inspecting power lines, specifically designed to monitor the gap between these lines and adjacent trees. The proposed inspection method comprises three primary components: data collection, extraction of the tree barrier, and distance assessment. Initially, an optical camera, affixed to the tower, captures two-dimensional (2D) imagery. Subsequently, a cutting-edge semantic segmentation algorithm, based on segmentation principles, is utilized to isolate both the tower and trees within these images. The final step involves applying a depth perception algorithm coupled with principles of geometrical optics to pinpoint the trees' location and accurately measure the 3D distances to the power lines. The experimental results indicate that this approach can precisely identify tree locations and measure distances. This innovative technique marks a substantial advancement in the monitoring of power lines, contributing significantly to the enhancement of safety measures for essential infrastructure.