Ice Thickness Identification of Transmission Lines Based on Aerial Images

In recent years, frequent extreme weather has led to increasingly severe icing of transmission lines, which can cause equipment damage and power outages. To address this problem, this paper introduces a method for identifying ice thickness on transmission lines based on aerial images. First, high-resolution aerial images are used to obtain real-time image data of transmission lines, and image preprocessing technology is used to improve the image quality. Then, convolutional neural networks (CNN) are used to extract features from aerial images and an ice coating recognition model is constructed. Subsequently, regression analysis is used to further optimize the prediction accuracy of ice thickness by combining the results of image processing and model training. Finally, the experimental results show that the model has high prediction accuracy in different ice thickness ranges, with MSE (Mean Squared Error) of 1.62 and R² of 0.97. In addition, in the test experiments of images with different resolutions, the MSE at ultra-high resolution is 1.98. Through hyperparameter optimization, the model achieved the best performance when the learning rate is 0.01 and the batch size is 32. In the above data conclusions, the method proposed in this paper can not only accurately identify the ice thickness of transmission lines but also has strong adaptability and practical application potential.