<p>With the development of UAV technology, UAVs play an important role in power transmission line inspection work. In recent years, solutions that synthesize tower coordinates and features to generate inspection routes have gained popularity. However, such solutions rely on a prior tower knowledge such as latitude, longitude, and altitude, even though such information may be difficult to obtain accurately. To address this situation, this report proposes an autonomous planning method for inspection routes based on target detection. (1) We propose a lightweight detector based on YOLOv8. (2) Based on the detection results, we design a path planner that contains three stages, to complete the initial tower information estimation and correction, as well as the subsequent towers information estimation. The designed Yolo model can be run at 56 fps on an onboard computer equipped with RK3588. Multiple flight experiments of continuous tower inspection also proved the effectiveness of the designed navigation information estimator, which provides an important reference for the development of power inspection technology.</p>