Abstract: Due to high degree of intelligentization, drone inspection of transmission lines is gradually becoming an important technical means of transmission line operation and maintenance, which can save time and effort. However, drone inspection has many problems, which cannot meet the needs of multi-element inspection of lines. Besides, it has poor real-time processing ability to detect major defects such as tree barriers and other hidden problems in real time. The degree of intelligentization is limited and can not meet the requirements of Unmanned Aerial Vehicle (UAV) operations in complex environments. Thus, a transmission line defect detection model is introduced for UAV inspection. The experimental results show that the equipment defect detector proposed has a good performance of equipment defect detection, for all types of defects with high comprehensive detection capability. The model has good adaptability to multi-scale targets and complex backgrounds, also can recognize the defective state of equipment.