Intelligent Bird’s Nest Hazard Detection of Transmission Line based on RetinaNet Model

Abstract

Electric energy is a kind of secondary energy with clean and efficient, convenient and safe use and the most wide range of applications. The transmission line equipment and the bird's nest on the tower will adversely affect the transmission line equipment, and even endanger the reliable operation of the transmission line. At present, the traditional manual line patrol has the problems of heavy workload, high cost, low efficiency and many blind areas, and the classical machine learning algorithm for bird nest recognition and classification is relatively low in efficiency and accuracy. In order to solve this problem, the RetinaNet model based on deep convolution neural network is selected for automatic detection of bird's nest targets. By adjusting the appropriate network structure and parameters to optimize the model, a RetinaNet model suitable for bird nest detection is established. The experimental results show that the accuracy of RetinaNet is 94.1%, and the recognition speed of each image is 68ms. Compared with Faster R-CNN, YOLO and SSD methods, the validity and reliability of RetinaNet model for bird's nest detection on transmission line equipment and towers are verified.