Bolts are important parts of transmission lines, and their states are closely related to the safe operation of transmission lines. Compared with insulators, U-shaped rings, and other fittings, bolts are small objects, and the detection of bolt and pin missing is difficult for transmission line patrol video analysis. This paper optimizes the parameters of the YOLOX network to adapt to the multiscale object detection task in the complex scene of the transmission line, and proposes a cascaded network model. The cascaded network locates the important large-scale fittings with bolts by the first layer and then detects the small-scale bolts on the fittings by the second layer, which greatly improves the detection accuracy of the small bolts. Finally, the effectiveness of the method is verified by experiments. The experimental results show that the cascaded YOLOX network can accurately detect the bolts that are originally difficult to detect. It effectively solves the problems of low detection rate of small bolts and pin missing defects.

The status of the insulators in power line can directly affect the reliability of the power transmission systems. Computer vision aided approaches have been widely applied in electric power systems. Inspecting the status of insulators from aerial images has been challenging due to the complex background and rapid view changing under different illumination conditions. In this paper, we propose a novel approach to inspect the insulators with Deep Convolutional Neural Networks (CNN). A CNN model with multi-patch feature extraction method is applied to represent the status of insulators and a Support Vector Machine (SVM) is trained based on these features. A thorough evaluation is conducted on our insulator status dataset of six classes from real inspection videos. The experimental results show that a pre-trained model for classification is more accurate than the shallow features by hand-crafted. Our approach achieves 98.7095% mean Average Precision (mAP) in status classification. We also study the behavior of the neural activations of the convolutional layers. Different results vary with different fully connected layers, and interesting findings are discussed.