Abstract: During the refined inspection of transmission towers by unmanned aerial vehicles (UAVs), a dedicated viewpoint is typically planned for each inspection target on the transmission towers, which requires the UAV to fly over a large number of viewpoints, resulting in low inspection efficiency. This paper proposes a UAV inspection route planning method based on multi-directional observation viewpoints to improve inspection efficiency by enabling a single viewpoint to cover multiple inspection targets. Firstly, initial unidirectional viewpoints for each inspection target are generated via sampling, and an optimized set of unidirectional viewpoints with wide visual coverage and high potential for multi-directional observation is obtained. Subsequently, based on the membership degree between each viewpoint and inspection target, unidirectional viewpoints are extended to correspond to multiple inspection targets. Infeasible observation directions are eliminated subject to the constraints of observation angles and gimbal angles, so as to acquire multi-directional viewpoints capable of monitoring multiple devices. Finally, an optimization model for UAV inspection route is constructed based on the set of multi-directional viewpoints. The Generalized Lin-Kernighan-Helsgaun algorithm is employed to determine the optimal visiting sequence of the viewpoints, and the A* algorithm is subsequently utilized to generate the final optimal flight route. Experimental results indicate that compared with routes planned with unidirectional viewpoints, the adoption of multi-directional viewpoints reduces the inspection route length for drum-shaped towers by approximately 14%.