Computer Vision-Based Damage Identification and Structural Analysis of a Practical Transmission Tower

For the purpose of field inspection and maintenance of transmission towers, computer vision-based damage identification and structural analysis method is proposed in this paper. Bending detection of tower members mainly includes the procedures of mask making, edge and contour detecting, and curving fitting. In the detection task of a damaged tower, the proposed method showed good robustness, and states of members were concluded according to Guide. Based on finite element analysis, a coefficient c j is defined as the product of state variability, risk of fail, and the relevance to the damaged element, which is able to describe association of members with the damaged one. Cases analysis with main member failing at different height showed that main members below cross arms, especially in the middle of tower have more associated elements, and should be paid more attention during inspection and maintenance.