With the rapid development of the economy, the scale of transmission networks has been expanding, bring higher demands and challenges on transmission line operation and maintenance. In this paper, the multi-rotor unmanned aerial vehicles (UAVs) safety inspection rules and automatic detailed inspection methods for transmission towers are studied. The theoretical model of multi-rotor UAVs of transmission lines inspection is established. The imaging calculation range of multi-rotor UAV-equipped cameras is determined. According to the inspection theory model, the flight safety judgment standard is designed, and the key parts of the transmission lines that need to be inspected are determined according to the accurate model of the transmission tower. Then the inspection waypoints are manually operated, and the photographing position and angle of each waypoint from the flight control are recorded through the waypoint planning function. Finally, the waypoints are connected in order, and the inspection route is automatically generated to achieve automatic detailed inspection. The inspection efficiency, error analysis and position accuracy comparing the proposed method with some state-of-the-art methods have been evaluated. The results show that the position error of UAVs automatic detailed inspection is less than 10 cm. The error of height is between 1.26 and 1.76 meter. Compared with the traditional manual inspection, the efficiency of multi-rotor UAVs automatic detailed inspection can be increased by 57.98%∼62.88% and can be applied to large-scale inspection of transmission lines. INDEX TERMS Transmission lines, unmanned aerial vehicles, inspection, automatic control.

Big data technology is more and more widely used in modern power systems. Efficient collection of big data such as equipment status, maintenance and grid operation in power systems, and data mining are the important research topics for big data application in smart grid. In this paper, the application of big data technology in fast image recognition of transmission towers which are obtained using fixed-wing unmanned aerial vehicle (UAV) by large range tilt photography are researched. A method that using fast region-based convolutional neural networks (Rcnn) convolutional architecture for fast feature embedding (Caffe) to get deep learning of the massive transmission tower image, extract the image characteristics of the tower, train the tower model, and quickly recognize transmission tower image to generate power lines is proposed. The case study shows that this method can be used in tree barrier modeling of transmission lines, which can replace artificial identification of transmission tower, to reduce the time required for tower identification and generating power line, and improve the efficiency of tree barrier modeling by around 14.2%.